Show Notes
Anthropic’s Claude was used in the military operation to kidnap president Maduro earlier this year. Why? Unclear. Was this legal? Absolutely not. More like this: AI In Gaza: Live from Mexico City
Surprise, surprise: the DoD feels that they should able to use AI models however they want, as long as its lawful — but… was this lawful? They are now threatening to designate Anthropic as a supply chain risk. What does this all mean? For this short, Alix was joined by Amos Toh, senior counsel at the Brennan Centre for Justice, to help us understand why the US defence department and an AI company are arguing about how best to us AI models for dehumanising and unjust military purposes. Further reading & resources: Pentagon's use of Claude during Maduro raid sparks Anthropic feud — Axios, Feb 13 Anthropic on shaky ground with Pentagon amid feud after Maduro raid — The Hill, Feb 19 US used Anthropic's Claude during the Venezuela raid, WSJ reports — Reuters, Feb 16 Pentagon Used Anthropic’s Claude in Maduro Venezuela Raid — WSJ, Feb 15 Amos’s Bluesky thread sharing more thoughts on the story Computer Says Maybe Shorts bring in experts to give their ten-minute take on recent news. If there’s ever a news story you think we should bring in expertise on for the show, please email [email protected] Post Production by Sarah Myles | Pre Production by Georgia Iacovou
More from Computer Says Maybe
The AI industry is so unpopular that it has started paying for the support it can’t earn. So what does it take for people to turn that mistrust into power that the industry can’t just buy off? More like this: Flock Shock: How Local Fights Over Surveillance Are Reshaping American Politics w/ Sarah Hamid & Anthony Ralphs
You may have seen the viral clip of a man in a full Darth Vader costume protesting Flock at a San Diego City Council meeting. The man behind the mask was Anthony Ralphs, who spent months trying to get the city to reconsider its contract with the surveillance tech company. After repeatedly raising concerns with the council, he turned to political theater - and got national attention. His protest is part of a growing backlash against Flock, bringing people across the political spectrum together to challenge surveillance tech and demand more say in how it’s used.
Tech companies keep telling us that the answer to solving climate change lies in AI. Meanwhile, the data centers required to power AI are keeping coal plants open longer, bringing in gas turbines, and driving plans for power stations bigger than anything we’ve built before. More like this: The Colossus Crossroads: Fighting for the Future of Memphis w/ KeShaun Pearson
When a tech executive says AI will be doing your job next year, it sounds like a forecast. Dr. Carissa Véliz argues it's something else: a command, delivered in advance, that works only if enough of us obey it. If you believe the prediction, you buy the product, reorganize your company, and bet the economy on it, and the prediction comes true because you made it so. So who gets to make predictions that stick? And what would it take to stop treating them as news and start hearing them as orders? More like this: AGI is Scientifically Impossible w/ Adam Becker In this episode, Alix Dunn talks with Carissa Véliz, associate professor at Oxford's Institute for Ethics in AI and author of Prophecy: Prediction, Power, and the Fight for the Future, From Ancient Oracles to AI. Carissa explains why a prediction is a command disguised as a description, how powerful people use algorithms to fulfill their own forecasts, and why the alternative just might be being relying on preparation over prediction. Further reading & resources: Prophecy: Prediction, Power, and the Fight for the Future, From Ancient Oracles to AI — Carissa Véliz Beware the Power of Prediction — Carissa Véliz, TED Talk The Origins of Totalitarianism — Hannah Arendt The Dangerous Ideas of “Longtermism” and “Existential Risk” — Émile P. Torres, Current Affairs
Almost everyone studying kids’ safety online agrees that a blanket social media ban won't work. And nearly every government moving on the issue is passing one anyway. This week, a researcher and a policy lawyer dig into why - and what could actually work instead. More like this: When Kids Can’t Consent: The Case Against Tech in the Classroom w/ Andrew Liddell