It’s been a whirlwind week in AI news. Weeks after a series of hacking incidents by rogue agents at OpenAI and Anthropic, we’ve seen a cascade of top AI researchers blow the whistle on what they call the catastrophic and even existential risks posed by the current pace of development. Then, over the weekend, Dario Amodei called for a slowdown in AI research — a call that was echoed by his competitors Sam Altman and Elon Musk.

If you’re feeling both concerned and hopeful at this moment, you’re not alone. Real threats are opening the door for real change. Tristan and Aza are going out into the world, talking to the media, technologists, and policymakers to turn this momentum into action.

Today on the show, Tristan shares how he’s feeling in this critical moment, breaks down the headlines from an insider's perspective, and points to tangible steps we can take right now to avoid the worst-case scenario.

You probably have questions about what’s happening. Good news: Tristan and Aza are gearing up for their annual Ask Us Anything episode. Pull out your phone, record your question, and send it to undivided@humanetech.com.

CORRECTIONS

The quote from Ajeya Cotra that Tristan cites is from her personal Substack, not the official METR report. She is also one of three contributors to the report, not the sole author.

Tristan incorrectly referred to Dario Amodei’s warning that an AI swarm could “take down” the internet. His actual phrasing was “take over” the internet.

President Trump’s call-in to the All In Summit occurred on Monday, 9/14, not Sunday, 9/13.

RECOMMENDED MEDIA

Watch “The AI Doc” on Netflix

Sign The Pro-Human AI Declaration

METR’s investigation of the Hugging Face incident

Dario Amodei’s call to “Pace the Frontier”

The Pacing the Frontier Open Letter

“On the Loose” by Dean Ball

"An Alien Mind" by Jakub Pachoki

"AI May Become the Third Superpower" by Paul Tudor Jones

RECOMMENDED YUA EPISODES

“Rogue AI” Used to be a Science Fiction Trope. Not Anymore.

The Self-Preserving Machine: Why AI Learns to Deceive

Daniel Kokotajlo Forecasts the End of Human Dominance

Former OpenAI Engineer William Saunders on Silence, Safety, and the Right to Warn

Mustafa Suleyman Says We Need to Contain AI. How Do We Do It?

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In a 2018 TED Talk, Yuval Harari argued that democracy prevailed over fascism and communism in the 20th century not because of any inherent advantage, but because the technology of that age favored open and democratic societies. It was simply too difficult and too costly to centralize power using pre-digital technology.

But, Yuval warned, with the rise of AI, the calculus might flip: it might actually become far more efficient to run highly centralized authoritarian governments.

Eight years later, we're starting to see what Yuval was warning about. Right now, police across all levels of government, from small-town sheriffs to massive federal agencies, are adopting AI at a staggering rate and using it to build the kind of efficient, centralized surveillance-state apparatus that Yuval warned about.

Take, for example, the recent deployment of hundreds of thousands of automated license plate readers (ALPRs) across the country by companies like Flock and Axon. These cameras have made headlines in recent weeks, but as you’ll hear in this episode, they are just the most visible frontier of an AI-powered digital dragnet that gives police the ability to analyze extraordinary amounts of data about private citizens.

Today on the show, Tristan speaks with two experts on how police are using AI technology today, what we can expect in the near future, and what protections we need to create now before it's too late.

Andrew Guthrie Ferguson is a law professor at George Washington University who studies how policing technology collides with constitutional law. He's the author of Your Data Will Be Used Against You: Policing in the Age of Self-Surveillance.

Garance Burke is an investigative journalist at The Associated Press, where she leads high-impact projects on the intersection of technology and society. Her investigations into the global surveillance industry were part of a package that won the 2026 Pulitzer Prize for International Reporting.

RECOMMENDED MEDIA

Your Data Will Be Used Against You by Andrew Guthrie Ferguson

Yuval Noah Harari’s 2018 TED talk: “Why fascism is so tempting -- and how your data could power it”

The full series of AP stories produced by Grance Burke and her Pulitzer Prize-winning colleagues

Further reading on Estonia’s data tracker tool

RECOMMENDED YUA EPISODES

Why Are Migrants Becoming AI Test Subjects? With Petra Molnar

America and China Are Racing to Different AI Futures

Laughing at Power: A Troublemaker’s Guide to Changing Tech

Corrections

Andrew said that real-time crime centers are in 300 cities across the US, which matches the findings of the National Real Time Crime Center Association; however, this figure is disputed, with the Electronic Frontier Foundation putting it at 150.

Andrew incorrectly described the resolution of the case of a man misidentified by AI in New Orleans. He said the man was acquitted, but in fact the charges were dropped.

Clarifications

As noted in the episode, Andrew’s discussion of agentic AI was a theoretical account of a tool that police may use in the future, not something actively used on the ground today.

Garance referred to the protests in Ferguson as the impetus for police-worn body cameras. Body cameras were being piloted in the US and UK prior to Ferguson, but federal grants in the wake of the protests made them much more common.

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In AI, what gets measured gets optimized. Right now, we're spending all our efforts to measure how capable and powerful models are, narrowly optimizing for those metrics while ignoring downstream consequences.

What if we could flip this dynamic on its head? What if, instead of what AI can do, we start to measure what AI does to us? What if, instead of races to the bottom on capabilities and engagement, we could incentivize races to the top on safety, or better yet, on making us more resilient and developed human beings?

That’s the mission of CHT’s Humane Evals program: we're bringing together researchers, psychologists, engineers, and technologists from across the entire AI ecosystem and beyond to build out the expertise and infrastructure we need to measure AI's impact on humans.

Today on the show, Aza Raskin explores the Humane Evals project with Imran Khan, a researcher and strategist who's been leading CHT's efforts in this area, and Jared Moore, a computer scientist and researcher who's been at the forefront of measuring AI's psychological impact on users.

If this sounds like something you're interested in working on, you can email us at evals@humanetech.com.

RECOMMENDED MEDIA

You can read more about Jared’s work at his website.

Related pieces by Imran on the CHT Substack:

• What is AI doing to humans? Why aren’t we measuring it?
• Will Human-Like AI Hijack Human Psychology?
• A Black Box Problem: Missing Data On How AI Affects the Human Mind

The website for the UC Berkeley Center for the Science of Psychedelics

KoraBench, the child safety AI benchmark that Imran referenced

RECOMMENDED YUA EPISODES

The AI Dilemma

Attachment Hacking and the Rise of AI Psychosis

How OpenAI's ChatGPT Guided a Teen to His Death

Corrections:

Aza gave the wrong year for Sewell Setzer’s death. It was in 2024, not 2025.

Aza referred to Joseph Henrich as an evolutionary psychologist; he was actually a biological anthropologist.

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.