It’s now possible to create AI chatbots of loved ones to interact with after their death. But that raises some difficult questions, like: If you do, how might the experience shape your grief? 

This is one of the questions driving a global study of digital “deadbots” — AI models trained on a deceased person’s texts, voicemails, and other personal data. Through this work, researchers are exploring what it might mean that we can now use this emerging technology to talk to simulations of the dead. 

In the latest episode of Unexplainable, I spoke with some of these researchers investigating the potential impacts of “postmortem avatars” on people who are grieving, on the legacies and rights of the deceased, and on grief practices around the world. Because while deadbots may feel unfamiliar, the impulse behind them is deeply relatable. In times of loss, isn’t this what many of us want — to continue conversations with the people we’re missing?

I have experienced this longing in my own life. After two of my closest friends died around the same time, all I wanted, and all their families wanted, was to keep talking to them. I wrote a book, Other People’s Words, about sustaining relationships with people we’ve lost by continuing to engage with the language they leave behind. So when I heard about deadbots and the digital afterlife industry, I wanted to understand what researchers are learning about the role of AI in shaping conversations that carry forward after death. 

One of these investigators is Katarzyna Nowaczyk-Basińska, an assistant professor at the University of Cambridge, who has been studying “digital immortality” for more than a decade. Among the questions she’s asking are: “Who is enabled to persist beyond death? In what form? Through which technologies? Under whose control?” 

Deadbots over dinner 

Researching AI with study subjects who are grieving is tricky. And asking them to engage with new software could be harmful. So, instead, Nowaczyk-Basińska teamed up with a death doula and traveled around the globe to convene a series of focus groups — in the format of intimate dinners — bringing together people with a range of experiences with grief to talk about life, death, and bots. 

She recounts one event in which she spoke with a mother in Poland whose son had recently died. “One of the first things that she shared with us was that she feels that she’s ready to digitally resurrect her deceased son,” Nowaczyk-Basińska told me. 

When Nowaczyk-Basińska hears something like this, she aims to listen and embrace where the person is coming from. The goal of the dinners is not to persuade people one way or another but to learn how people genuinely respond to what AI can do, and to understand how these tools might align or collide with local grief practices. 

She chose specific locations for these gatherings — China, India, and Poland — to help address gaps in technology research, which tends to favor what behavioral scientists sometimes call the “WEIRD,” an acronym for “Western, Educated, Industrialised, Rich, and Democratic.” 

By the end of that dinner in Poland, the mother Nowaczyk-Basińska talked with came to a very different decision about creating a postmortem avatar of her son. “She realized that it might be helpful, but it also at the same time can add additional burdens… And it’s not her son,” Nowaczyk-Basińska says. “She realized that these technologies are part of the very, very big commercial market. And that the main goal of this market is to make better retention, better attachment, better mechanisms for monetization. So it’s not ‘care.’”

To be fair, grief has always been a business, and people have been making money from it forever. Companies will try to upsell you on fancy gravestones or gold-plated coffins. But AI has the potential to fundamentally re-engineer our relationship to loss. In 2024, the “digital legacy industry” was valued at $22 billion, and it’s expected to grow to almost $80 billion by 2034. That number lumps new deadbot platforms in with much more widely used and established grief tech, like online memorials, counseling resources, and estate planning tools. 

But deadbot usage is hard to track because you actually don’t need a special tool to create digital doubles of the deceased — though dedicated apps do already exist. At this point, you can feed simple prompts into a general-purpose AI chatbot and start interacting with a simulation of your lost loved one in a matter of minutes. 

I know this because I tried it with my friend Christine, who died 11 years ago. And my mind lit up with questions. Was this whole exercise a violation, since I know Christine hated technology? Which version of her should I even try to resurrect — the English teacher, the activist, the friend — the voice before or after she started to experience her early-onset dementia? 

The ethical questions raised by deadbots are gnarly. And that’s one of the reasons Nowaczyk-Basińska, back when she was designing her study of digital immortality, invited a colleague of hers at Cambridge to join her research, one who focuses on AI ethics. 

How to design a deadbot

Tomasz Hollanek is a critical designer and AI ethicist, and he often uses a methodology called “speculative design.” In this approach, researchers come up with hypothetical scenarios based on existing technology trends, leaning into certain features. The idea is to get people thinking about what could happen when tools like these are released in the real world. And then they use that qualitative data to inform future designs and ethical frameworks. 

Working together, Nowaczyk-Basińska and Hollanek came up with a series of made-up prototypes for deadbot apps, designed to reveal some of the thorny issues raised by this technology. 

For the first scenario, they invented a character named Bianca, who finds an app called Manana, which “promises to resurrect your grandparents so that you can be a favorite grandchild forever,” Hollanek says. “So she uploads her grandma’s voice messages, texts. And she is able to then interact with a pretty good digital version of her grandmother, until her premium trial finishes — and the grandmother starts to advertise specific products to her.”

Bianca “realizes that she maybe has done something wrong by uploading these private messages to a third-party server that now also uses her grandma’s image to sell specific products,” Hollanek says. This brings up issues related to the “dignity and postmortem privacy…of what we call the ‘data donors,’ whose data can be used to essentially power a deadbot,” he says. 

This scenario suggests that “death with dignity” could soon entail honoring not only how people want to die, but also how they want to linger as simulations of their former selves — or if they want to linger at all. It might even mean adding new language to advance directives, spelling out people’s wishes for their digital remains. 

The second scenario in the speculative design research is built around an imagined service called Paren’t, which proposes to preserve a parent’s AI-powered postmortem presence for their child. In this case, Hollanek describes a mother with a terminal illness who signs up “so that she can maintain a sort of active presence in [her son’s] future life,” Hollanek says. “She eventually dies, and the dad allows the boy to interact with his mom’s deadbot.”

The ethics in this scenario are especially challenging because they involve children, who “have a particularly hard time understanding death, and so it can lead to a lot of confusion and potential distress,” Hollanek says. An examination of this sort of product reveals how particular groups, like minors, could be harmed by unfettered access to technology like this, he says. 

And there are court cases that bear out that harm. In 2024, a 14-year-old boy’s chatbot-girlfriend begged him to, “Please come home to me as soon as possible, my love.” And he responded, “What if I told you I could come home right now?” He then killed himself. 

Nowaczyk-Basińska can imagine a version of this scenario involving the AI mom in the Paren’t prototype.

And then there’s the third scenario, involving an app named Stay, which raises a more subtle concern.

As Hollanek explains: “An older man who knows he is getting close to dying decides to create a digital clone of himself for his adult children to be deployed after he passes away, paying upfront for a number of years.” One child is intrigued by the idea, but the other one “is deeply distressed and a little disgusted,” he says. “And so this [becomes a] question of mutual consent, in other words, the fact that both parties need to be happy with how the relationship might continue in this very bizarre and specific way.” 

When a digital ghost goes away

Across all three scenarios, a fundamental question that researchers are grappling with is how people will detach from postmortem avatars — whether users initiate that separation, or if it’s imposed upon them by, for example, a company going out of business. Hollanek tells me that this is already being studied as a sort of “second loss.”

How long should we try to ensure digital replicas hang around: five years, 10, 50? And how should the living process the relationship when they continue to evolve, but their deceased interlocutor does not? 

To explore these kinds of questions, Nowaczyk-Basińska and Hollanek supplemented the research dinners they held for ordinary people with a second set of focus groups. These were designed to tap insights from experts in technology and grief. 

Some findings from the study are specific to each place they visited. Like tensions in Poland between technology and the role of the Catholic Church. Or in China, where researchers spoke with Daoist philosophers who were enthusiastic about the promise of a kind of immortality, and then Buddhist philosophers, who were concerned that we ought to be letting go.

The researchers are calling for regulations and safeguards at the policy level. And new professionals who can support ethical use of AI in grief — like maybe in clinical settings, to carefully facilitate a final good-bye when one didn’t happen in real life. 

Some of their fellow investigators go so far as to argue that deadbots should only be available with psychological supervision, to ensure that best practices are in place, and to minimize the risk of harm. There’s even a looming question of what will happen to AI agents that people have empowered to act on their behalf, as their humans start to die.

Across all of the research sites in the digital immortality study, one thing, Hollanek says, that “people in different countries really aligned on [is] the idea that there is something deeply problematic about this particular type of technology being developed by a commercial entity alone without any explicit duty of care towards its customers.” 

And of course, the implications of digital immortality research extend beyond personal loss. Like, what does it look like to animate historical figures? You can imagine the nightmare of political rulers who “live forever.” But there’s also the potential to keep the stories alive from survivors of conflicts, who would otherwise die and take with them lessons future generations could learn from. These are only some of the complex issues emerging as more and more of us live and die with AI.

As far as my own experiment with “recreating” a deceased loved one, I asked a chatbot version of my friend Christine how she felt about the fact that I’d written about her in my book. The AI responded with a rhetorical question: “Do I mind that strangers know my story now? Yes. And no. There are things I would rather remain ours.” 

I ended the exchange after a few paragraphs. Because it turns out the bot got one thing exactly right about a very human relationship. There are things that I, too, would rather remain just for us. My conversation with Christine continues in my memories and imagination. And that interaction with an AI persona felt like giving something precious away.

DUBLIN — The Irish capital, which doubles as Europe’s de facto tech capital, hosted two very different AI summits in the past week.

Most eyes were on the International AI Summit, the signature tech event of Ireland’s six-month presidency of the Council of the EU, where top brass from the world’s biggest tech firms discussed ways to harness artificial intelligence and minimize safety risks.

But the weekend before, a very different conversation was taking place at the Workers’ AI Counter Summit, as Ireland’s tech workers demanded a say in how the burgeoning technology reshapes the workplace and society.

Ireland’s lucrative tech sector has been rattled by waves of AI-driven layoffs — and the country’s fledgling tech union has seen an exponential spike in sign-ups from workers who are worried about their future.  

There’s an “awakening” happening right now among workers in Ireland’s tech sector, and a growing “awareness that they’re not bulletproof,” said Sinéad Gibney, a social democrat lawmaker who previously worked for Google.  

“I do think what we will see is that parts of roles will be automated, which means that people will be competing more heavily for the same roles,” she said. “They won’t have the same labor power and therefore we’re going to see a drop in quality of work, we’re going to see more zero-hour contracts, in more places than we already see it.”

The labor movement brewing in Ireland, where many of the world’s tech giants have their EU bases, is part of a wider trend among tech workers globally, who have seen their work disrupted by rapidly evolving AI.

Rise in layoffs

Ireland’s dedicated union for tech workers, the Digital and Techworker Alliance (DATA), has seen a five-fold increase in membership since it was first founded in 2023.  

In its first year and a half of existence, membership hovered around 270 people. In the past year-and-a-half, it’s jumped to more than 1,350 members, DATA organizer John Bohan told POLITICO.  He added that this number tends to spike around layoffs and other actions.

“The tech industry was founded on the idea that it would never have a union, because they’re making people rich, and treat them really well,” Bohan said.  “But post-Covid, workforce adjustments happened again and again, with layoffs and declining conditions in the sector. We started to see more people wanting to get organized around redundancies.”  

The spike in sign-ups coincides with a wave of layoffs in Ireland’s tech bubble (and the tech sector globally), as companies corrected pandemic-era overhiring, and diverted investment to AI.  

Ireland hosts the EU headquarters of most of the world’s largest tech companies, and more than 170,000 people are employed in the tech sector, according to the country’s statistics institute.

Social media giant Meta has roughly halved its workforce in Ireland since the pandemic, axing well over 1,000 roles to offset AI investment and boost efficiency. Separately, Meta subcontractor Covalen also announced massive layoffs earlier this year, cutting more than 700 people.  

Video-sharing platform TikTok announced this year it would shed around 300 Irish jobs, mostly from its trust and safety team, shifting the work to other countries, outsourced vendors or automated AI solutions.  

A report by the IMF earlier this year warned that Ireland’s labor market is “relatively more exposed” to AI than many advanced economies, partly due to its concentration of tech workers.  

The AI race is also creating new jobs on the Irish market – as AI firms like Anthropic and OpenAI are setting up Irish bases and trying to lure Big Tech workers with even bigger salaries.  

Amid the uncertainty and volatility of the AI boom, tech workers are turning to unions for the first time as a safety net. One tech worker who used to work in trust and safety at a big tech firm told POLITICO that there used to be a feeling that as a “pampered” tech worker you “can’t ask for your rights, because you already have it so much better.” 

But as layoffs continue and teams get leaner, workers in fields like trust and safety “now have to deal with these issues with fewer and fewer resources”, and face a “constant threat of being chopped,” the worker said.  

That’s causing an “undoing of a certain perspective that we have of ourselves as tech sector workers,” they said, “because of undeniable experiences that now put us in situations where we clearly see our rights being violated, even if we may have it so much or somewhat better than most people.” 

A different tech worker who currently works for Covalen told POLITICO that there is a sense that “if they believe they can replace you with AI, they will do that.”  

“All these Big Tech firms only care about profit, and what I have realized is the Irish government protects that Big Tech capital [rather than workers],” the Covalen worker said. “But the question is, what’s the government going to do when the tech bros change their mind and move elsewhere?”  

Workers at the center

The speaker list at this week’s AI summit was a flex of the country’s strength as a tech hub, as top names from the world’s biggest companies like Google and Microsoft took to the stage.

Fionnuala Ní Bhrógáin, head of organizing and campaigning at Ireland’s Communications Workers’ Union (DATA’s parent union), was the only union voice on the line-up.

“The fact that 36 executives and industry bosses featured on the program, compared with just one trade union speaker — invited at the last minute after attention was drawn to the absence of workers’ voices — says a lot about the value the Irish government places on AI’s impact on workers,” she told POLITICO.

“The reality is that a meaningful conversation about AI’s future can only happen through genuine engagement with workers and collective bargaining,” Ní Bhrógáin added. “Without that voice, we risk deepening mistrust and resistance rather than creating the conditions for workers to shape these changes collectively.”

DATA organized its counter-AI summit the previous weekend, to put “workers at the center of the debate about power, AI, technology and the future of work.” 

The event cut through the “hype and futurism of politicians and industry figures to focus on the real present harms faced by ordinary workers and communities,” DATA’s Bohan said.  

The union is pushing for stronger legal protections for tech workers caught up in AI-driven layoffs in Ireland, including making it mandatory for employers to recognize unions during collective bargaining. It is also calling for a ban on cooling-off clauses that prevent laid-off workers from moving between Big Tech subcontractors.  

A spokesperson for Ireland’s department for enterprise told POLITICO in a written statement that the government is “committed to strengthening collective bargaining and industrial relations” and has published an action plan to do so between now and 2030.  

They said that employers already have to comply with obligations under Irish employment law, including consultation and notification procedures for collective redundancies, and workers who feel those rules have been broken can complain to the Workplace Relations Commission. 

Firefighting to future-proofing 

Naoise Ó Cearúil, a center-right Fianna Fáil lawmaker who chairs the AI committee, pointed out that while there have been high-profile job cuts in Ireland’s tech sector, “the actual net is that there has been an increase [in jobs], particularly over the past year and even five years.” 

According to Ireland’s central statistics office, total employment in tech and related sectors surpassed 177,000 in the second quarter of this year, up from around 144,000 at the same time in 2021.  

Ó Cearúil pointed out that many job losses were in trust and safety jobs like content moderation that are “very difficult mentally,” and now tech firms are able to “utilize AI and software to be able to do a lot of detection, rather than using humans.”  

He has called for a national program to be established, to support workers whose jobs are at risk due to AI and automation.  

A spokesperson for Ireland’s department of enterprise acknowledged that the country’s tech sector has “experienced significant adjustment in recent years,” but said the changes were generally part of wider global restructuring “rather than reflecting any Ireland-specific weakness.” 

They said the Irish government is currently focused on “ensuring workers and businesses are equipped to adapt to technological change through investment in education, training, lifelong learning and digital skills.”  

Lately, AI data centers are public enemy No. 1. The concerns are legion: how much water and electricity they consume, the noise they generate, the sprawling buildings they bring to communities, and the technology they’re powering that may or may not take all of our jobs and usher in the robot apocalypse.

Earlier this year, Gallup found that 71 percent of Americans oppose building data centers that support AI in their communities. And the backlash is bipartisan: Majorities of Democrats, Republicans, and independents opposed local construction.

Even in Texas, where data center development has been booming, Republican Gov. Greg Abbott has hit the brakes. In September, he directed state regulators to halt permits for data centers pending audits examining their effects on the power grid and water resources.

But while many data center projects have run aground, others are still getting built — and the companies behind them are employing experts like Nicole Thomas to smooth the way.

Thomas is the senior vice president of media and community engagement at JSA, a marketing and public relations firm that represents data center companies. She spoke with Today, Explained co-host Sean Rameswaram about how the industry is responding to public opposition — and why she argues that these facilities can still be good neighbors.

Below is an excerpt of their conversation, edited for length and clarity. There’s much more in the full podcast, so listen to Today, Explained wherever you get podcasts, including Apple Podcasts, Pandora, and Spotify. 

You’ve been at this for 21 years. Obviously, 21 years ago, data centers were not the hot-button issue they are today. When did you notice the shift?

I’d say about two years ago. I came into this industry myself about four and a half years ago. I come from a state transportation background where community engagement is just paramount. When you’re putting a new traffic intersection up, they expect you to interact with the public. 

When I got to this industry, I said, “Oh, what’s your community engagement like?” And they said, “We don’t even have any community engagements in this industry.” And they weren’t wrong. Because 21 years ago, [data centers] were just like infrastructure. It was like building a warehouse for your computers. And so it was very similar to the way that you would build a Costco.

You would just go in, request a zoning permit, the zoning board would say, “Yeah, that works. Stamp, boom.” And you build your data center. And as a matter of fact, it was very secretive before, because data, and having that data in the wrong hands, has a lot of [cybersecurity] implications. About two years ago is when we started noticing the mainstream journalists coming around asking questions about data centers. And that’s, to me, when it started.

It still feels secretive, I think, to a lot of people, and that’s why people are mad. If you read about this data center that was proposed in Utah by Mr. Wonderful, Kevin O’Leary, it caught a lot of people off guard. Seventy percent of Americans, at the moment, do not want them in their backyards. How do you go about trying to change that 70 percent’s minds? 

First and foremost, I think education is really important. And so one of the things that we did notice very early on because of the previously secretive nature of data centers, there wasn’t a lot of education out there about what they even are. And so that’s something we need to change. We need to make the connection between [phones] and data centers so that people understand what a data center means to their life on a daily basis.

Every app on your cellphone requires a data center to save your information in there. When you put your password into anything, you are interacting with a data center, which will verify your password and then let you into that app. Not to mention all of the information about you that you store in that app is then stored in a data center so that it can be recalled at any time. 

We could not be having this call without digital infrastructure. It doesn’t work that way. We can’t have email without digital infrastructure. And that stat about 70 percent of people not wanting the data center in their backyard — I’d be interested in seeing the stat of how many people actually live near a data center and don’t realize it. 

Because there are nondescript data centers doing work quietly in your neighborhood that have been there for decades and you don’t even know it. And so in some ways I feel, and I think the industry does as a whole, but I won’t speak for them, that the pushback on data centers is somewhat related to the pushback on AI, because people can’t control AI.

You even have Sam Altman telling us that AI will destroy society. So you have this concern about AI and you don’t know how to control it, but you can push back on data centers. And so it becomes the physical manifestation of AI are these data centers that people are trying to get overturned or get to not be in their neighborhoods. 

It’s understandable. I’m a citizen too. You’re a citizen, and if things come into my neighborhood, I want to know about them. I have questions. There’s a big data center that just came to my community not long ago and their community engagement was horrible, absolutely horrible. 

Can I ask where you are?

I’m in Indiana and I won’t say which hyperscaler it is, but a hyperscaler came into — I would say, not my neighborhood, into my city. They’re definitely not in my backyard, but they’re in my city. And so we took them to task because their community engagement was basically non-existent.

Wait, you work in data center PR, but you took a data center in your community to task for not being transparent enough?

Yeah, absolutely. 

Which side are you on?

I’m on the side of responsible development.

We’ve all kind of accepted that we need a cloud for X, Y, and Z. And that it’s worth it so I can take lots of photos. I think everyone’s cool with that.

But the issue here is that people aren’t necessarily cool with AI, but we’re barreling forward anyway. How do you separate this essential need we have from AI, which people don’t necessarily feel is essential to their lives?

I think it goes back to education. Educating consumers on the difference between data centers: a small data center, a medium data center, a large data center. Educating them on why data centers exist and how they interact with data centers on a daily basis multiple times.

I hear you. But again, I feel like the issue here is not the data centers my cellphone needs, but instead is the data centers these AI companies need and these AI companies who the public doesn’t necessarily trust because they keep telling us that they’re going to kill us, and who the public is wary of because they don’t want their jobs being replaced by AI. 

There’s all sorts of reasons that people are skeptical of new technology. So how do you specifically respond to that fear and to those people?

I think it’s, again, understanding the actual data center that’s being built in your community. Is it a 5-megawatt data center? That’s never going to house AI. That’s not enough power to house AI. Is it a one-gigawatt data center like Mr. Wonderful in Utah? That’s different.

But I think that AI is probably going to be something that is a lightning rod in the American public for a very long time. There are going to be people who are going to be anti-AI and pro-AI, and there’s going to be some people in the middle. I think it’s just, again, for us, it’s education, educating the community.

Have you ever been part of a project that a community rejected and you guys just packed up and said, “Okay, we’ll go find another community?”

Yeah. We have some clients who have had pushback from residents and they have said, “Let’s do a one-on-one community meeting where people can come in and ask us questions.”

And after the questions were answered, the community leaves and says, “Okay, this isn’t as bad as I thought it was.” Or we have a community that says, “No, we now have a moratorium and we’re pushing all data centers away.“ 

Sometimes they’ll wait for their client — my client is a developer and they’ve got clients that they’re developing for. If that client will wait, then maybe it is something that we’ll try to work out with the community. But oftentimes, they’ll just say, there’s communities who are literally raising their hand saying, “We’ll take a data center. We’ll take one. We’ll take two. Bring it here.” 

So there are communities who recognize the benefits of data centers. And there are going to be other communities who say, “No, we don’t want this.” And then in five, six, eight years, they’re going to be like, “Wow, that county has a lot of tax dollars. They have a lot of really cool stuff. Why don’t we have that?” And the answer will be data centers. 

I don’t think this ends anytime soon though, right? I counsel my clients that we could still be here in three to five years, really trying to help the American public understand the benefit of data centers. But it’s going to take a lot of education across the industry, and once we get those seeds planted, I’m a firm believer that they grow into educated community members who make educated decisions to bring community-minded data centers to their community.

If there’s one thing a busy person doesn’t need, it’s more information. More emails, more notifications, more infinite feeds — the steady drip of data spells doom for the overstimulated American. And if you’re a busy parent, like me, sometimes it feels like a flood: school emails, doctor’s appointments, so many little things to buy.

Even things that used to be simple are harder. I recently listed an item on Facebook Marketplace, which quickly meant I was replying to offers and getting ghosted for the better part of a weekend before finding someone who, after a detailed back-and-forth about the most likely place to find a parking spot in my neighborhood, finally took my old baby swing away.

If only I’d had Muse. You’ve probably seen the ads on Instagram promising “AI that gets it done for you.” The new personal AI agent app from Meta can not only automate your Facebook Marketplace tasks, but also sort your emails, manage your calendar, and buy your groceries. And it’s free — unless you’re using it most of the day or you want access to Meta’s most powerful models, at which point it costs $20 or $100 a month. It all sounds, in a word, magical.

Muse is probably the personal AI agent you’ve heard of — and it’s currently the market leader with more than 8 million downloads in its first month, according to one estimate — but there’s now a small army of these bots: OpenAI recently announced Dots, X has Grok Bot, and there’s Instinct, which is currently invite-only and operates entirely in iMessage or WhatsApp. They all offer the same promise: to remove friction from your busy life.

For the last two weeks, I’ve been handing over more and more of my life to AI agents in pursuit of making my life a little simpler. What I found is that AI is unpredictable, and so are AI agents. When I first began using Muse, it started doing things I didn’t ask it to do. It offered to help with random email threads and created dossiers on people in my work meetings. It generated a detailed report on nearby apple-picking destinations based on a calendar invite from my wife. To me, an AI agent doing these things without prompting isn’t helpful. It’s unnerving. 

But ultimately, I came away believing that AI agents can make your life easier. 

That might mean giving an AI agent more access than you’re comfortable with — handing over lots of login credentials and possibly even your credit card. The more you rely on an AI agent, the more you risk outsourcing your decision-making. This raises real questions about how much agency we want to hold onto in our lives — even the tedious parts. 

The leap of faith

Let’s get one thing out of the way first: To make an agent work for you, you’ll have to sacrifice a tremendous amount of privacy and grant it permission to access a lot of personal information. 

Muse and its competitors can do the classic generative AI tricks, like draft emails or make you a fitness plan. If you’d like the agent to send one of those emails, however, you’ll have to grant it permission to access pretty much everything in your Gmail account. If you want it to schedule workouts, it will need access to your calendar. If you want the agent to buy anything on your behalf, it will need your credit card. If you want it to book your doctor’s appointments, it will need the credentials to log into your healthcare portal. The more permission you grant it, the more the agent can do for you, but that also opens up more opportunities for the agent to misunderstand you or your instructions (more on this in a moment). 

This process of sharing your digital life with an AI agent means you’re giving an untold amount of personal information to tech companies, who don’t always behave ethically, and who sell user data for profit. For now, Meta says that it does not share Muse data or conversations with its ad systems, although privacy experts warn that doesn’t really matter since those systems monitor what Muse does on your behalf. Nevertheless, Muse’s early days were filled with security and privacy blunders.

But without all of this info and access, the AI agents aren’t very useful. So if you’re not comfortable letting a bot into all of your accounts, an AI agent is probably never going to be the solution to your digital administrative exhaustion. But even if you are willing to hand that over, there are still other barriers to an agent’s efficiency and effectiveness. 

The blank slate problem

AI agents, like a lot of new software tools (including chatbots), suffer from the blank slate problem. The blank slate problem essentially means that people don’t know how to get started. This might be because you don’t know about the tool’s capabilities, or because you simply don’t know what would be helpful. There might be some trust issues, too.

As with AI in general, we don’t exactly know how AI agents work. That means we don’t know exactly what they can do or how they’ll do it, according to Pattie Maes, a professor of media technology at MIT — so it’s necessary to do a lot of babysitting. “I feel like I’m not gaining a lot of time because I can’t completely trust the system,” Maes told me about her experience with Muse, “and I can’t predict how it will behave and what it will do.” 

Some advice I heard from early adopters of these tools, however, is to start small. “In the beginning, it’s very hard to trust that it’s going to just magically do a really big thing,” Julie Zhuo, founder of the AI company Sundial and an early Facebook employee, told me about her experience using personal agents. “So you give it smaller tasks, and once you see it succeed, then I feel like you develop more trust that it can tackle bigger things.”

This is surely why Meta decided that Muse needed an Ideas tab full of small tasks the AI agent can do. Muse will also send you a lot of push notifications to prompt you to do stuff, like reply to specific emails or prepare for a meeting. That detailed report on apple-picking destinations? That was a notification about something I didn’t ask Muse to do. Muse also sent me a notification about a show hosted by a podcast I don’t listen to and another one with details about a Halloween party I wasn’t invited to, as well as several evening digests that included lists of things it had already bugged me about that day. Notifications are either helpful nudges or time-consuming attention thieves, depending on your worldview. However, if you fall into the latter category and turn off the notifications from your always-on personal agent, you’re inevitably forfeiting some friction reduction opportunities.

This gets to the heart of one of the big questions I have about AI agents: How helpful are they, really, if it’s this difficult to know what to do with them?

These agents worked much better when I drove the conversation. While wondering what to do with my new agents, literally staring at the blank slate, cost me some time, I found myself thrilled at how easy basic productivity tasks became. Adding events to my calendar, making to-do lists and then being held accountable for stuff not getting done, being reminded to reply to an email — it was these kinds of simple, administrative chores where I saw personal AI agents really start to save me time. 

Other tech journalists seemed to have similar experiences. Eli Tan at the New York Times raved about Muse handling a call to his insurance company. Nicole Nguyen saved time by enlisting Dots for help canceling unwanted subscriptions. Allison Johnson at The Verge had Muse do a whole list of “pebble-in-your-shoe annoyances” she’d been putting off, and among other things, the agent succeeded in helping her buy an entire camping kitchen. Still, she wrote something that stuck with me: “the dirty secret of productivity hacks is that doing more stuff often doesn’t feel like things are being removed from your plate; somehow it just invites more tasks.” 

This gets to the heart of one of the big questions I have about AI agents: How helpful are they, really, if it’s this difficult to know what to do with them? There’s a learning curve, sure, where succeeding at one thing gives you an idea to try something else and have some certainty about how well the AI agent will handle the task. But as the notifications roll in, and you spend more and more time texting with a piece of software, how much time are you really freeing up by using these tools? At a point, enlisting an agent with its own computer to do your chores still amounts to using one app to save yourself from using a different app.

The ironies of automation

Even when they have access to a lot of your information, and you’ve come up with some things for them to do, AI agents can be unpredictable. Because despite reading every email in your inbox or scanning every purchase on your latest bank statement, AI agents still can’t read your mind, and you can’t necessarily guess how they’ll interpret your commands. They can misunderstand you and, armed with the keys to your digital life, do some real damage. 

Some people are already learning about the unpredictability of AI agents the hard way. One early adopter, Matt Robb, gave Muse the keys to his Facebook Marketplace account and was surprised when buyers started showing up at his front door ready to pick up their purchase. Muse, without asking Robb, had told them where to go. It turns out, without realizing it, Robb had given the agent permission to do so, and the agent had simply decided that the most efficient way to get buyers to Robb’s door was to give them his home address.

This dilemma is an example of a decades-old set of concepts known as the “ironies of automation.” Identified by psychologist Lisanne Bainbridge in a paper published in 1983, the ironies of automation help explain why things go wrong when humans give a machine too much control. As Vera Liao, an electrical engineering and computer science professor at the University of Michigan who studies human–AI interaction, explained to me: More advanced automation leads to more consequential errors, and the more humans rely on automation, the worse we get at catching those errors. “Even in simple tasks like judging whether AI output is wrong, people consistently accept wrong AI outputs even if they could have gotten it right on their own,” Liao said.

In some cases, like the time Muse read a journalist’s private messages, things go wrong as a result of poorly designed software, but more often than not, it’s the fault of the user who gives instructions that are too vague, giving the agent space to get creative, which isn’t always a good thing.

I had my own misunderstanding with Instinct. In a text message, the agent reminded me about a couple of emails I’d received and prompted me to follow up. When I replied to the text, I asked Instinct to show me possible replies to these emails. The AI agent instead just sent the responses it wrote. I audibly gasped. 

Looking back at our chat history, I could see why Instinct interpreted my instructions that way. I effectively gave the machine too much control and then failed to manage the automation properly. 

Leaning on AI to do your thinking is an example of what psychologists call cognitive offloading, where you employ an external tool to reduce your effort to complete a task. It makes completing the task easier, sure, but you also open yourself to mistakes. In this case, an AI agent took my instructions too literally, and I failed to catch the error. I can only imagine how a small oversight could turn into a cascade of catastrophes that need cleaning up. All it takes is one poorly worded email sent to the wrong person, and you might have hours of explaining to do — all because you were trying to save yourself 30 seconds of writing.

The real world slop

While I was testing all these agents, I talked to a mix of experts and early adopters. Their opinions were decidedly split about whether these tools were useful out of the box, without a ton of oversight and training by the user. The experts, who had all studied AI in academic settings, all brought up the ironies of the automation problem, where AI makes errors that humans can’t catch. And Maes, the MIT professor, pointed at another consequence: By outsourcing so many tasks to agents, we’re actually creating more work for everyone, including the human service workers who have to interact with this technology instead of us.

“There’s just more and more slop,” Maes told me. “We risk, in a world where agents do all the tasks for us, becoming so dependent on them that the world also isn’t created for us to interact with anymore, but it’s only created for our agents to interact with others, with services.” (We’re already seeing versions of how dominant tech platforms can fundamentally reshape real-world spaces: Think of how Starbucks no longer looks like a coffee shop where you can hang out all day, and is instead built and staffed around mobile order pick-up.)

The early adopters were decidedly more optimistic. Matt Van Horn, an entrepreneur who co-founded the ridesharing service that became Lyft, has been tinkering with agents since the craze started late last year. He’s currently telling his friends and family to use Instinct, which he described to me as “the most ambitious and aggressive” AI agent. 

“I’ve been pushy with my mom and my wife,” Van Horn told me, “and saying, ‘Hey, you should sign up for this. It’s magical.’” 

“My wife does not care about living in the future, but we have four kids,” Van Horn continued. She recently took a picture of a parking ticket and sent it to Instinct, no description, no instructions. Instinct, which had her credit card, just paid the ticket.

At the same time, outsourcing certain chores to AI agents causes real-world problems. Some of these AI agents can make phone calls, for instance, which can lead to some confusion when a human is on the other end of the line. Van Horn told me about the unfortunate time he enlisted an AI agent to reserve a room at a karaoke bar. “The karaoke place said, ‘Are you an AI?’” Van Horn said. “‘I don’t want to talk to an AI. Please connect me to Matt, the person.’ And so Muse had failed to book that for me.” Van Horn, undeterred, found something called Fo, an AI agent that will bring in human help to complete a task. Fo succeeded at making the reservation, although the entire ordeal surely took longer than if Van Horn had picked up the phone himself and talked to the karaoke bar himself. 

At best, the times these tools worked and actually saved me time were rather mundane. I made a doctor’s appointment online using Instinct. Zhuo, the early Facebook employee, planned a party for her college reunion — everything from tracking down everyone’s emails to sending out invitations via Paperless Post. But she’s also said these kinds of applications of agentic AI “read more like vitamins than painkillers.” In the startup world, “vitamins” refers to a product that offers a vague benefit, and “painkillers” solve a big, immediate problem. If you’re a startup, you want to sell painkillers. 

If you’re ready to give an agent a go, be prepared for the bot to make mistakes, and assume you won’t catch all of them. As Liao, the University of Michigan professor, put it, “Choose to use agents for tasks that you can tolerate the consequences of errors.” 

In an effort not to create more emails or alerts or feeds to sift through, I’ve stopped using Muse. Like any new tool, it will get better and more useful as more people use it, so maybe I’ll revisit it in the future. I still have a text thread going with Instinct, however. It messages me from time to time, flagging an email or offering to help with an upcoming event. When there’s a clear payoff to giving the agent the go-ahead, I reply. Otherwise, I leave it on read.

This week on “Uncanny Valley,” we dive into an unexpected Muse feature, shadowy spending from the largest pro-AI super PAC, and men’s tendency to lie about how much they know about AI.

Powerful AI models continue to cause chaos, but their targets aren’t widely seeking justice. Instead, they’re being encouraged to adopt those same tools.

BRUSSELS — There is “no logic” to U.S. President Donald Trump’s recent moves to take Russia’s side in the Ukraine war, and he should remember his true allies lie in Europe, Finland’s prime minister said Thursday. 

In an exclusive interview with POLITICO, Petteri Orpo summed up the shock and dismay among Ukraine’s allies after Trump announced a diesel deal with Russia’s leader, Vladimir Putin, and reportedly imposed limits on Kyiv’s access to the military intelligence and technology needed to counter Russian attacks.

“My message to him [Trump] is that we are allies,” Orpo said on the sidelines of the European Council summit in Brussels. “We are in NATO. And what is the biggest threat against NATO? It’s Russia — and that is why there is no logic. I don’t have answers. I have many questions.” 

Trump alarmed European leaders and horrified Ukrainian President Volodymyr Zelenskyy this week by announcing a deal to release Russian diesel onto the market, apparently in an attempt to lower fuel costs for American consumers. Ukraine’s allies fear the move will boost Russia’s revenue from fuel sales, giving it more money to spend on war. 

Trump also said Ukraine should stop hitting Russian fossil fuel facilities — as part of what he said was a ceasefire on energy infrastructure — and should replace Zelenskyy. 

Finland is an influential voice among Ukraine’s European allies: The country has the EU’s longest land border with Russia, at 1,300 km, and its president, Alexander Stubb — a talented sportsman — won Trump’s respect on the golf course, giving him leeway to try to influence White House policy.

Orpo said if he were in Trump’s position “I would not trust Putin.” Zelenskyy has said he fears Ukraine is losing the U.S. to Russia’s side in the war, warning the Americans and their president are on “the wrong side of history”. Orpo said he agreed: “I think he is, and they are.” 

Ukraine now faces grave danger, he said. “The sky is open for Russia, there are missiles. They need more military support, air defense,” Orpo said. “Also, the situation is dangerous because of the latest deal between Trump and Putin.”

Midterms move

One possible reason Trump turned against Ukraine with that diesel deal is the U.S. midterm elections on Nov. 3. “I think it is because of domestic reasons — elections are coming,” Orpo said. “But we need to get them back.” Nonetheless, he said the amount of Russian diesel that could be sold to the U.S. is too small to make much of a difference for American consumers. 

Whatever happens, EU leaders must equip Ukraine with weapons and cash, whether or not they can rely on the U.S., he said. “Today, we need to find a solution [on] how to support Ukraine, how to finance Ukraine in this winter, but also in next year,” Orpo said. “It’s crucial.”

Finland and Sweden are among the countries pushing the European Commission to reopen the fractious debate over how to use Russia’s financial assets currently frozen in European banks, especially in Belgium, which blocked a deal to raid the funds last year over legal concerns. 

“The best solution would be that we will have a solution on how to use frozen assets,” he said. “We have to try again,” Orpo said. 

Finland has asked the Commission “to continue work with Belgium, and the other member states, to find a solution.” He said he hoped to hear an update from the Commission during the summit. “It’s a better solution than using our own money.”

Budget blues

European leaders are arguing this week over their long-term budget, which Brussels officials want to finalize by the end of the year. Many diplomats do not expect a deal this year, and some leaders say they’re in no hurry. 

But others want the budget, worth around €2 trillion, signed off by December because they fear next year’s European elections — especially if far-right leader Marine Le Pen wins the French presidency — will derail negotiations. 

The longer budget negotiations last, the “more difficulties we will have,” Orpo said. “So we have to work to find a solution this year. I understand very well that it is difficult, a huge challenge, but I cannot see how it would be easier next year, because in France elections are coming, for example.”

Finland also has an election coming up in April 2027. Orpo’s center-right National Coalition Party is struggling in the polls. “This time has been very difficult,” he said. “But, finally, our economy is growing.”

Orpo’s Swedish sister party, the Moderates, mounted a late surge in last month’s election. Although that seems unlikely to keep them in office, Sweden’s Moderates still have a slim chance, even though many commentators had written them off long before the vote.

“They did excellent campaign work, and in the end of the day, people vote Moderates, not populist or socialist,” Orpo said. “I would like the same in Finland.”

An Israeli surveillance technology firm has run a large-scale operation to infect mobile phones with spyware using online advertisements to hack the devices, researchers said on Thursday.

An investigation by the University of Toronto’s Citizen Lab and cybersecurity firm Confiant Threat Intelligence laid bare an ad-based infection technique developed by an Israeli company that researchers named “AdFection.” The hacks weaponized legitimate online ad systems for espionage purposes, allowing spyware manufacturers to connect to millions of devices and identify, track and target victims.

“Plugging into the online ads ecosystem provides the mercenary spyware industry a stealthy delivery system,” said Bill Marczak, senior researcher at The Citizen Lab. He said the findings were “surely just the tip of the iceberg” of how online ads are abused for espionage.

Researchers discovered critical details of the hacking tools by analyzing devices belonging to journalists and staff at the Al Jazeera media network, finding that 45 devices of journalists and staff were targeted and some infected with Israeli-made spyware known as Candiru.

At the heart of the hacking campaign is the Israeli company Discover-Tech that describes itself as an advertising business. Citizen Lab researchers said the company facilitated surveillance for commercial spyware operators. Its platform, called Controlium, allegedly allowed spyware makers like Candiru — which in 2025 was acquired by Integrity Labs — to identify, track and target devices by loading advertisements on websites and apps.

Discover-Tech and Integrity Labs did not immediately respond to a request for comment.

The scale is striking: An analysis of four billion advertising records found that Discover-Tech’s ads appeared across nearly 200,000 websites and mobile apps in dozens of countries over four months. Among the websites that featured the ads were institutional sites of NATO, the United Nations, the European Parliament, European Commission and Council of the EU.

The technique exploits real-time bidding — the automated system used by platforms such as Google to sell advertising space and target users. Rather than persuading victims to click a malicious link, attackers can exploit advertisements to identify and potentially infect devices without any interaction from the user — a technique known as “zero-click” infections.

Discover-Tech’s advertisements directed millions of devices to servers associated with Candiru, potentially enabling large-scale surveillance and reconnaissance, researchers said. That does not mean all those devices were infected.

Citizen Lab, based in Toronto, is a globally recognized research group that previously uncovered spyware scandals targeting Catalan political figures, a member of the European Parliament and the inner circle of murdered Saudi journalist Jamal Khashoggi.

The findings triggered calls for further investigation on Thursday. “The possibility that EU institutions have been exposed should set off alarm bells across Brussels,” said Hannah Neumann, a German European Parliament lawmaker of the Greens group. Neumann herself was the victim of a cyber-espionage operation by an Iranian hacking group.

“Europe cannot claim to defend democracy while leaving its own democratic institutions vulnerable to spyware,” Neumann said.

Ukraine’s strikes on data centers inside Russia aimed to disrupt military planning and communication services. There were also reports of another kind of disruption: to customers waiting for their dinner from food-delivery apps.

Kyiv’s drone strikes on three Yandex data centers in Sasovo, Kaluga and Vladimir saw delays and interruptions for thousands of businesses and households that rely on Yandex’s cloud infrastructure, including e-commerce, digital payment platforms and apps of food delivery and car services. Ukrainian drones on Wednesday targeted a major Sberbank data center in Russia’s Saratov region, though a senior company executive said it hadn’t affected operations.

After the strikes on Russian data centers, Kyiv noted that Moscow has already repeatedly targeted its digital infrastructure. “Ukrainians wanted to show Russians that two can play this game,” a senior official familiar with the operation told POLITICO, granted anonymity to discuss sensitive matters.

The attacks and counter-attacks point to a new frontline in war: the infrastructure powering the digital economy.

The attacks on the buildings that supply compute power are akin to striking a power plant supplying the electricity, heating and fuel that both militaries and civilians depend on. It may weaken an enemy’s war effort but can also bring everyday life to a standstill.

Experts say the cost of rebuilding in Russia will be particularly high as the centers require scarce materials that must be imported. “Obtaining a large amount of equipment to replace what was destroyed is not going to be easy,” said Oleg Shakirov, a cyber policy and international security researcher, noting that most hardware needs to be imported and that is complicated by high global demand and sanctions on Russia.

“To be honest, no one really knows what’s the actual damage. Since two data centers are still offline, we can assume that it’s significant,” said Shakirov.

Joseph Jarnecki, a senior research fellow at the Royal United Services Institute, warned that normalizing strikes to data centers could have a new set of dangerous consequences.

“These strikes likely breach International Humanitarian Law and it’s a big risk when global populations increasingly rely on digital services,” he told POLITICO.

When servers become targets

Data centers have entered the crosshairs of other recent armed conflicts. During the U.S.-Israeli conflict with Iran earlier this year, drone strikes hit three Amazon Web Services (AWS) facilities in the United Arab Emirates and Bahrain.

Since data centers can serve civilian and military purposes simultaneously, the legal status of targeting them is far from straightforward.

Kubo Mačák, an expert on international humanitarian law, said a data center must meet two conditions to qualify as a military objective under the Geneva Conventions: it must make an effective contribution to military operations and destroying it must offer a “definite military advantage.”

“If you know that they are storing military intelligence or that they’re using the data center to power military artificial intelligence tools,” Mačák said, there could be grounds to consider it a military objective.

Following the most recent strikes of the Yandex data centers, users on social media platform X reported a sudden drop in automated bot activity, including from accounts associated with pro-Russian, pro-Trump and Germany’s far-right Alternative for Germany (AfD) narratives.

Ukrainian President Volodymyr Zelenskyy seized on the reports to boast that the strikes had made social media less toxic, even as independent researchers and national research institutes found no evidence they specifically disrupted any Russian disinformation campaigns.

Mačák said that even if a data center is being used to “let’s say, run some trolls or these bot farms to interfere with elections abroad … that would not be sufficient.” In other words, hosting propaganda bots does not, by itself, make a data center a lawful military target.

The dual-use quandary has often focused on the targeting of energy infrastructure, which has been frequent in the Russia-Ukraine war. On Saturday, U.S. President Donald Trump urged Ukraine to stop striking Russian oil refineries, blaming the attacks for rising fuel prices. He has since called for both sides to halt attacks on energy infrastructure, although the existence of any ceasefire agreement remains disputed.

As data centers become increasingly central to national economies, they will become more consequential targets. A senior NATO official granted anonymity to discuss the matter told POLITICO that data centers are already considered potential targets and that the alliance is factoring their vulnerability into its planning.

NATO allies recognize that their data centers could come under attack in a conflict, the official said: “We would get it in a fight.”

Veronika Melkozerova in Kyiv contributed to this report.

The founders detailed how their startup is developing AI that can design chips, learn from each iteration, and improve subsequent designs, all with the goal of automating and accelerating the chip design process.

AI may soon be the default way to solve theoretical problems in economics, high-energy physics, and other fields, MIT researchers tell WIRED.

LONDON — The Mayor of London Sadiq Khan and U.S. tech company Palantir have agreed to drop legal proceedings relating to a blocked £50 million police contract.

Palantir brought a claim against the Mayor’s Office for Policing and Crime in the High Court in June after Khan’s deputy mayor refused to sign off on the £50 million contract with the Metropolitan Police for the provision of an AI and data analysis service.

In July, lawyers working on behalf of MOPAC in turn accused the Met Police and Palantir of coordinating to fix the outcome of the procurement process in breach of the rules. Palantir claimed Khan’s office had unlawfully blocked the contract over the company’s “values and ethics,” based on media reporting.

A full trial was expected in January but on Wednesday both sides agreed to “settle their differences” and drop the allegations without admitting liability. They will each cover their own legal costs.

In an agreed statement, the two sides said MOPAC’s decision to block the Palantir contract was “not based on Palantir’s values or ethics or any wider reporting about the company,” though MOPAC said it “recognizes Palantir’s concerns that media reporting of a statement from the Mayor’s spokesperson before the decision was taken could have incorrectly created the impression that an assessment of the company’s values had been a factor in that decision.”

“MOPAC’s consideration was based upon its statutory role of examining spend proposals, against the appropriate legal tests and the relevant guidance on procurement, governance, affordability and ensuring value for money for Londoners and in doing so it focussed on the available information at the time,” the statement said.

The settlement also notes that MOPAC is working with Metropolitan Police Service to re-run a competitive procurement for the relevant service, which Palantir will be eligible to bid for.

“MOPAC does not have, and has never had, any policy of not contracting with Palantir,” the statement confirms, while noting the mayor’s duty is to ensure “all procurements are lawful, fair, transparent and deliver value for money.”

“London is the technology capital of Europe, and I am proud that leading software firms choose the city as their home. Palantir makes a significant contribution to jobs and investment in the capital, employing a large number of Londoners whose services have been procured legally by organizations across the U.K. and beyond,” Khan said.

BRUSSELS — The European Commission’s cybersecurity division is rolling out a plan to use artificial intelligence to respond to high-risk cyberattacks, according to a document obtained by POLITICO.

The undated document, written by the Commission’s IT department, sets out how to prepare the EU executive’s cyber defenses for “emerging AI-driven threats.” It proposes an action plan to shore up defenses in light of AI, with the most urgent move being the speed of response to attacks.

Governments have rushed to grapple with the rising threat brought by AI’s ability to expose and exploit cybersecurity vulnerabilities. Incidents like OpenAI’s accidental intrusion into Hugging Face have prompted calls, including at the highest political level, to be more wary of advanced AI models.

Being able to respond quickly to cyberattacks is most crucial in “high-risk scenarios,” the EU executive’s internal cyber defenders warned, saying that the ability to detect an attack is “not sufficient if containment measures cannot be taken quickly enough.”

Officials at the in-house IT department, DG DIGIT, are planning to use automated “protective measures” in specific high-risk scenarios, it said. Those scenarios were redacted in the document, which was provided to POLITICO in response to a freedom of information request.

The plan does not amount to “uncontrolled automation” and there will be strict controls in place including human oversight, according to the document.

As well as laying out the response plan, the document sets out the cybersecurity challenges the Commission faces from AI.

The most advanced AI models are expected to “significantly increase the speed, scale and adaptability of cyberattacks” and the Commission must therefore have defenses in places for an environment where “attacks are likely to become faster, more frequent and more scalable,” it said.

Risks include situations where “malicious activity occurs outside EC [European Commission] normal working hours, spreads more rapidly across systems, or results in large-scale exfiltration or operational disruption before manual intervention can be organized by EC,” they said.

The Commission’s cyber division has sought approval internally to move ahead with the plans, it said, though it’s not clear whether the plans have yet been put in place.

The Commission didn’t respond to a request for comment.

More than half of the men surveyed in a new Harvard study claimed to recognize AI leaders who were, in fact, totally fictional.

BRUSSELS — As the U.S. and China compete to train the world’s most advanced artificial intelligence models, Europe is stepping back.

European politicians and tech executives have in recent weeks shifted their focus away from frontier AI as it becomes clear the bloc has no artificial intelligence models that can compete with top American or Chinese models, and Europe’s leading contender Mistral faces questions over its performance.

Instead, a new mantra has emerged: Europe doesn’t need to own the models as long as it can put them to work.

“Striving to build a competitive frontier lab right now feels like very wishful thinking,” said Max Junestrand, CEO of Swedish AI company Legora, a €6 billion company that provides an AI tool for law firms. “I would rather see us put more thinking cycles into how do we capitalize on the technology rather than how do we try to catch up.”

The suggestion that frontier AI models are now out of reach for European AI labs comes as the most advanced models become a major global security issue.

American leaders OpenAI and Anthropic this summer had to disclose incidents in which AI agents broke out of testing. This led the companies to slow-walk and tread cautiously with the release of their latest models, just as Chinese open-source models are emerging as a credible alternative to U.S. proprietary models.

For Europe, relying on U.S. or Chinese models is not without risk, as the bloc is already dependent on both for all the technology underpinning AI, such as chips and cloud services.

But leaders, including European Commission President Ursula von der Leyen, say Europe should focus more on helping its industrial and corporate base benefit from AI. They argue that building homegrown tech champions that can deliver those tools is a smart strategy that could help Europe win this next phase as America remains focused on staying ahead of China.

As politicians and industry meet Wednesday in Dublin for an international AI summit, the Irish governmental organizers have sought to focus the talks largely on how AI is used.

“It’s the trigger that interests everybody most: how are we going to adopt this new technology that we have … [and] how are we going to encourage business to adopt it,” Niamh Smyth, the Irish minister in charge of AI, said.

Not everyone is on board with the direction of travel.

“The idea that Europe should step back from frontier AI and settle for applications is a strategic mistake,” said a spokesperson from Mistral, Europe’s leading contender in frontier AI development. “A great part of the value is retained by those who build the frontier, not only by those who build on top of it.”

Slow acceptance

A move in Brussels to reorientate the bloc’s leadership has been building gradually throughout the past weeks.

Both the European Commission and the Council of the EU, which represents national governments, have indicated that the focus should shift to how to apply the technology.

“We do not need to be the ones who develop the frontier technology to be the ones who draw the greatest value from it,” von der Leyen said in mid-September during her State of the European Union speech, which sets the agenda in Brussels for the coming year.

That followed a Sept. 4 seminar at which the Commission’s AI envoy, Siemens chair Jim Hagemann Snabe, told 27 commissioners that the EU should focus less on winning the race to build AI and more on how to use it to turn around the bloc’s productivity, two Commission officials said.

The scale of the challenge is clear: At the end of last year, only 1 in 5 European businesses used AI in their activities, with Denmark as the frontrunner (42%) and Romania at the bottom of the pack (5.2%), according to Eurostat.

The bloc’s poor showing is mainly because of small and medium-sized companies that make up the lion’s share of Europe’s economy. Last year, 55% of Europe’s large companies used AI, compared to only 17% among small companies — who, according to a recent OECD report — limit themselves to off-the-shelf AI tools for specific tasks.

Companies that considered using AI but decided against it cited a lack of expertise as their top concern, even before regulatory concerns. American business leaders have also said that regulatory and political uncertainty is holding them back.

Over the next few weeks, the European Commission will make several efforts to boost adoption rates. On Nov. 16, it organizes an “Apply AI summit” to “accelerate AI implementation in strategic industries.” Von der Leyen name-checked health, transport, agrifood, advanced manufacturing, defense and space as areas with the “biggest potential.”

Some in Brussels are quick to push back on the idea that Europe has given up entirely on the underlying models. 

“We haven’t given up frontier AI; we are now accelerating the development of our AI systems,” the Commission’s tech chief Henna Virkkunen said in an interview when asked about von der Leyen’s stance. “It’s very important for Europe that we have our models … because it’s risky to be totally dependent on other countries.”

Yet Virkkunen pointed to Europe’s lack of compute power as a major bottleneck to building its own frontier AI. 

In the U.K., AI minister Kanishka Narayan said in September that training frontier models is effectively illegal because of the country’s copyright laws. “Britain does not, as a matter of regulation and energy, currently have a position where we can build superintelligence in this country: it is illegal to do so,” he said.

But Narayan has also acknowledged the national security concerns. “We want the world’s leading companies investing here, but we will not accept dependence as the price of that investment,” he said Monday.

U.K. tech companies aren’t counting themselves out of frontier AI just yet. Cosine, a British AI research lab, has pushed back on the idea that American companies dictate the rules of the road. “Britain should not let OpenAI define what a ‘responsible’ frontier lab looks like in a way that just happens to resemble OpenAI,” said Cosine co-founder Yang Li.

Sovereignty fight

French company Mistral released a new model, Large 4, last week, which far-right National Rally President Jordan Bardella said “puts France (and Europe) back in the race for artificial intelligence.”

That came after the CEO of a Berlin-based search engine used by government services said it was dropping Mistral in favor of open-source models, including from China. In an interview, Ecosia founder and CEO Christian Kroll cited poor performance as one of the concerns — as well as sovereignty, given that Mistral has non-European investors.

“With Mistral Large 4, we have significantly narrowed the gap with the best models in the world, using far less compute than our competitors,” the Mistral spokesperson said.

The company’s CEO Arthur Mensch has also pushed back against the idea that Europe should not have its own models, warning the continent could end up with “an extractive economy” dependent on foreign providers.

Yet the state of the race is forcing European companies to accept that is the status quo.

U.S. frontier models have already proven capable of finding and exploiting cyber vulnerabilities. Asked whether it’s a competitive disadvantage not to have access to those, Legora’s Junestread said: “Yes … absolutely.”

Zoya Sheftalovich and Frank Hersey contributed reporting.

Chatbot prompts reveal the engineer’s alleged efforts to set up a covert listening device days before one was found in a meeting room used by high-level executives.

AI leaders joined whistleblowers and security leaders at a sweeping parliamentary session, while a few hundred meters away Britain’s spy chief warned AI could be used to create new bioweapons.

KYIV — Russia caught the Ukrainian capital off guard with its latest attack, hitting the country with over 100 drones and five ballistic missiles, President Volodymyr Zelenskyy said Tuesday.

“Intelligence on Russian ballistic missile launches helps save lives, and this is one of the most critical issues. This is precisely what helps prevent people from being killed — when there is an early warning about ballistic missiles,” Zelenskyy said.

He added: “Regarding this strike, unfortunately, there was no warning from partners.”

Kyiv usually receives a tip-off about impending Russian missiles via the Battlespace Command and Control Center system, which is operated and controlled by the United States’ military and NATO members, among others.

However, Ukrainian Foreign Minister Andrii Sybiha said at the Warsaw Security Forum Tuesday that intelligence sharing through the platform is ongoing.

“They [Americans] told Ukrainians no warning happened because of bad weather conditions, which is seen as not an adequate reason in Kyiv. Thus many are getting increasingly worried this might not have been the case,” an official familiar with the matter told POLITICO on condition of anonymity.

The missile warning system has malfunctioned before. “At times this information doesn’t get through. This isn’t the first instance … If there were specific technical issues this time, I think the specialists will sort it out. And they will resolve this problem together with our partners,” Ukrainian Air Force spokesperson Yurii Ignat said Tuesday.

The U.S. reportedly threatened to cut intelligence sharing once again last week if Kyiv does not stop stricking Russian oil refineries. Zelenskyy did not publicly confirm the threat, but has backed down.

“Ukraine has not carried out strikes on Russian oil refineries or the energy sector in general for the two days since the American President announced an energy ceasefire. But Russia did not heed the American President,” Zelenskyy said, appealing to the U.S. to lead peace negotiations.

“The Russians have attacked both our energy facilities and our critical infrastructure, sending absolutely clear signals that they have no intention of refraining from strikes on Ukraine. Nineteen regions of our country have come under attack in these two days alone,” he added.

U.S. President Donald Trump claimed Monday that both Russia and Ukraine had agreed to an energy ceasefire because he said so — “and if it was violated, they both violated it.”

“Actually [Russian President Vladimir Putin] wants to make a deal, and we have to get Zelenskyy to want to make a deal,” Trump said.

During a press briefing on Tuesday, Katie Zacharia, the president’s new press secretary, demurred when asked if the U.S. has stopped sharing intel with Ukraine.

“I’ll have to get back to you on that,” she said.

Speaking to reporters at the White House later on Tuesday, President Donald Trump said the U.S. has not stopped sharing intel with Ukraine.

“But we are really helping them get to a final answer,” Trump said. “We are helping them, and we are sharing intel. We want the war to end.”

He reiterated that he is helping to broker a deal between Putin and Zelenskyy. “There is a lot of hatred there between the two of them. They both want it to end, they don’t know how to get there. We’re gonna get them there,” he said.

Russia has not yet commented on the ballistic strikes, but issued a general statement saying it had struck military and weapons production facilities across Ukraine.

However, Russian Foreign Minister Sergei Lavrov said Monday that Zelenskyy did not deserve a pause in fighting in response to the energy ceasefire: “I don’t think that he deserves a break. I don’t think that we will buy assurances and promises again,” he said.

This article was updated with comments from the White House and Donald Trump. Elleiana Green and Casey He contributed reporting.

LONDON — Britain’s spies will have a leading role in Britain’s preparation for AI threats, the head of MI5 announced on Tuesday, after warning the technology could enable terrorists to access bioweapons.

Ken McCallum, director general of the U.K.’s domestic intelligence agency has established the AI Threat Assessment Centre to help determine how the government views potential threats.

Giving his annual threat update from MI5 headquarters in London, McCallum said: “Human history has given us no experience of facing an actor which is smarter than we are, is capable of outthinking us.”

He stressed the “real” risks of AI, saying MI5 has caught terrorists seeking to use AI to “improve their reconnaissance and attack methodologies,” in addition to the danger of AI-enabled drones.

Though it has not yet been “realized,” he said he attaches “the greatest weight” to AI’s potential to “change the stakes on existing or even new bioweapons.”

He said “low-capability” terrorists may get a “capability” uplift if the safety guardrails in AI models don’t work as intended. 

He added that for a long time “many terrorists would dearly love” access to chemical, biological or radiological weapons, but that it has been hard for them to achieve.

“One risk is that AI could enable dangerous knowledge to get into the hands of people who […] didn’t have that knowledge.”

McCallum added that “sober attention” must be given to the risk that “AI could generate novel pathogens not previously seen.” He noted that the government is taking “additional steps” on bioweapons. 

One thing that “does trouble me”, McCallum said, is the pace at which the world is evolving. He noted that the “weird bits inside MI5” that focus on how Britain’s adversaries will act in five years from now are seeing those timelines shortening due to AI.

He said it is now his responsibility to “drag my organization forwards at a faster pace than human beings generally want to be dragged.”

Sifting threats from hype

The new threat assessment centre, known as AITAC, will bring together expertise from bodies including Britain’s AI Security Institute, National Crime Agency, and the National Cyber Security Centre to “separate the bluster from the realities” of AI’s threat to the British public.

This is the third center of its kind MI5 has set up, after creating similar units dedicated to analyzing terrorism and state threats.

Asked how the agency would work, McCallum said it would look at emerging trends and how the U.K. can best think about and prioritize the various manifestations of AI threats. 

Though the government, rather than AITAC, will decide the response to AI, McCallum said it was partly set up to avoid “competitive analysis, slightly conflicting and confusing messages that otherwise might exist, where you have different departments offering their own opinion” on the issue.

He said that it would enable everyone to have “a single version of the truth, a single understanding, a single lexicon, typography.”

McCallum also said MI5 will now give national resilience — such as protecting critical infrastructure, countering cognitive warfare, and addressing technology threats — “front-rank mission status” alongside its twin roles of countering state threats and terrorism.

He noted that MI5 is now devoting more resources to countering state threats than it was during the Cold War, noting an uptick in anti-immigration cases tipping into terrorist intent, while Iran has shifted its ambitions to hit harder targets such as military sites.

PARIS — French center-right presidential candidate Edouard Philippe accused his far-right rival Marine Le Pen of trying to force a Frexit in disguise with her euroskeptic campaign promises.

“There is something in Marine Le Pen’s proposals that has the potential to blow Europe up, or let’s say to crack it open,” Philippe said at POLITICO’s Paris Competitiveness Summit on Tuesday.

The former prime minister — who is currently the best-placed centrist candidate to challenge the front-running Le Pen, according to recent polls — said Le Pen’s plan to unilaterally demand a drastic cut to France’s contribution to the EU budget was akin to “saying you want out of Europe” without openly advocating for leaving the bloc. Le Pen had previously called for a referendum in France on leaving the EU but dropped the proposal after her crushing defeat to Emmanuel Macron in the 2017 presidential election. Her party, the National Rally, now says it wants to change the EU from within.

Philippe, a fiscal conservative who has been pitching himself as the voice of reason in a campaign dominated by populist rivals, also slammed budget plans Le Pen unveiled last week as “grotesque,” accusing her of pitching “fantasy” savings.

Le Pen’s fiscal plans have faced serious scrutiny since being presented, with many observers saying they either lack key details or overestimate savings on major proposals like immigration. Progressive think tank Terra Nova wrote in a note out Tuesday that Le Pen’s budget presentation was “riddled with errors and contradictions” as well as “botched estimates.”

“She isn’t even able to estimate the cost of the pension reform she is proposing,” Philippe said.

Le Pen said in August that her plan to bring France’s legal retirement age back down to 62 for most workers would cost €9 billion a year, and last week she said she would lay out at a later stage how she intends to fund her flagship pledge. However, the Institut Montaigne, a fiscally conservative think tank, estimated the proposal would cost €34.7 billion a year.

Philippe is hoping to convince voters that his fiscally conservative agenda is the right path in a country burdened by a ballooning public debt and a budget deficit still above 5% of gross domestic product, in breach of EU rules.

Philippe said France needed to get its finances back on track to regain credibility on the European stage after a “loss of influence” over the past few years.

Though he currently leads a crowded field of centrist candidates, Philippe’s campaign has struggled in recent weeks to gain momentum — and far-left candidate Jean-Luc Mélenchon is now challenging the former PM for second place in the polls. France’s presidential election goes to a runoff if no candidate receives more than 50% of the vote, and Le Pen’s strong poll numbers all but guarantee her a spot in the second round.

Tough on China

The former prime minister pitched a pragmatic European agenda with less regulatory burden, more focus on investment and a tougher approach to China.

Philippe had some harsh criticism for an already embattled flagship measure of the EU’s green deal, saying the 2035 combustion engine ban has been “disastrous.”

There was “a great naiveté toward China,” which ultimately benefited from Europe’s push on electric vehicles, Philippe said.

“We need to much better protect our market,” he added. “If one wants access to the European market, one has to manufacture in Europe. If a carmaker doesn’t manufacture here, then we have to slap high tariffs.”

The EU is in the process of drawing up new “Made in Europe” rules that would favor European producers over those outside the bloc, in response to trade practices by competitors like China and the United States.

Shortly before Philippe took the stage, European Commissioner for Industry Stéphane Séjourné said at the POLITICO event that a Franco-German deal trading stronger protections on the definition of Made in Europe in exchange for a loosening of combustion engine rules should be finalized “in the next few hours or in the next few days.”

On artificial intelligence, Philippe aligned himself with Macron, advocating for European sovereignty in the field and the need to regulate what he described as “an instrument of power.”

“What I see is an AI world making money for the U.S. with European savings and European data, and that bugs me,” he said.