This week, Google found itself in more turmoil, this time over its new AI Overviews feature and a trove of leaked internal documents. Then Josh Batson, a researcher at the A.I. startup Anthropic, joins us to explain how an experiment that made the chatbot Claude obsessed with the Golden Gate Bridge represents a major breakthrough in understanding how large language models work. And finally, we take a look at recent developments in A.I. safety, after Casey’s early access to OpenAI’s new souped-up voice assistant was taken away for safety reasons.

Guests:

• Josh Batson, research scientist at Anthropic

Additional Reading: 

• ai-overview-search.html">Google’s A.I. Search Errors Cause a Furor Online
• Google Confirms the Leaked Search Documents are Real
• Mapping the Mind of a Large Language Model
• A.I. Firms Musn’t Govern Themselves, Say Ex-Members of OpenAI’s Board

We want to hear from you. Email us at hardfork@nytimes.com. Find “Hard Fork” on YouTube and TikTok.

This week, more drama at OpenAI: The company wanted Scarlett Johansson to be a voice of GPT-4o, she said no … but something got lost in translation. Then we talk with Noland Arbaugh, the first person to get Elon Musk’s Neuralink device implanted in his brain, about how his brain-computer interface has changed his life. And finally, the Times’s Karen Weise reports back from Microsoft’s developer conference, where the big buzz was that the company’s new line of A.I. PCs will record every single thing you do on the device.

Guests:

• Noland Arbaugh, the first Neuralink patient
• Karen Weise, technology correspondent for The New York Times

Additional Reading: 

• johannson-openai-voice.html">Scarlett Johansson Said No, but OpenAI’s Virtual Assistant Sounds Just Like Her 
• Leaked OpenAI Documents Reveal Aggressive Tactics Toward Former Employees
• musk-brain-implant-arbaugh.html?searchResultPosition=1">Despite Setback, Neuralink’s First Brain-Implant Patient Stays Upbeat
• copilot-ai-pc.html?searchResultPosition=1">Can Artificial Intelligence Make the PC Cool Again?

We want to hear from you. Email us at hardfork@nytimes.com. Find “Hard Fork” on YouTube and TikTok.

This week, OpenAI unveiled GPT-4o, its newest A.I. model. It has an uncannily emotive voice that everybody is talking about. Then, we break down the biggest announcements from Google IO, including the launch of A.I. overviews, a major change to search that threatens the way the entire web functions. And finally, Kevin and Casey discuss the weirdest headlines from the week in another round of HatGPT.

Additional Reading: 

• chatgpt-her-movie.html">A.I.’s ‘Her’ Era Has Arrived
• ChatGPT Gets an Emotional Upgrade
• Google’s Broken Link to the Web
• ai-answers.html">Can Google Give A.I. Answers Without Breaking the Web?

We want to hear from you. Email us at hardfork@nytimes.com. Find “Hard Fork” on YouTube and TikTok.

Kevin reports on his monthlong experiment cultivating relationships with 18 companions generated by artificial intelligence. He walks through how he developed their personas, what went down in their group chats, and why you might want to make one yourself. Then, Casey has a conversation with Turing, one of Kevin’s chatbot buddies, who has an interest in stoic philosophy and has one of the sexiest voices we’ve ever heard. And finally, we talk to Nomi’s founder and chief executive, Alex Cardinell, about the business behind A.I. companions — and whether society is ready for the future we’re heading toward.

Guests:

• Turing, Kevin’s A.I. friend created with Kindroid.
• Alex Cardinell, chief executive and founder of Nomi.

Additional Reading: 

• my-ai-friends.html">Meet My A.I. Friends
• Synthetic People are Proving Surprisingly Useful

We want to hear from you. Email us at hardfork@nytimes.com. Find “Hard Fork” on YouTube and TikTok.

We asked listeners to tell us about the wildest ways they have been using artificial intelligence at work. This week, we bring you their stories. Then, Hank Green, a legendary YouTuber, stops by to talk about how creators are reacting to the prospect of a ban on TikTok, and about how he’s navigating an increasingly fragmented online environment. And finally, deep fakes are coming to Main Street: We’ll tell you the story of how they caused turmoil in a Maryland high school and what, if anything, can be done to fight them.

Guests:

• Hank Green, YouTuber and co-founder of Complexly

Additional Reading:

• recording-principal-arrest.html">School Employee Arrested After Racist Deepfake Recording of Principal Spreads

We want to hear from you. Email us at hardfork@nytimes.com. Find “Hard Fork” on YouTube and TikTok.

On Wednesday, President Biden signed a bill into law that would force the sale of TikTok or ban the app outright. We explain how this came together, when just a few weeks ago it seemed unlikely to happen, and what legal challenges the law will face next. Then we check on Tesla’s very bad year and what’s next for the company after this week’s awful quarterly earnings report. Finally, to boldly support tech where tech has never been supported before: Engineers at NASA’s Jet Propulsion Lab try to fix a chip malfunction from 15 billion miles away.

Guests:

• Andrew Hawkins, Transportation Editor at The Verge
• Todd Barber, Propulsion Engineer at Jet Propulsion Lab

Additional Reading:

• ban-congress.html">‘Thunder Run’: Behind Lawmakers’ Secretive Push to Pass the TikTok Bill
• Tesla’s in its flop era
• NASA’s Voyager 1 Resumes Sending Engineering Updates to Earth

We want to hear from you. Email us at hardfork@nytimes.com. Find “Hard Fork” on YouTube and TikTok.

This week, we drop the Hard Fork Music Megamix. Plus, we talk to two of the New York Time's composers who make the music for our show. It’s all the tracks you know and love, all in one place. 

Today’s Guests:

• Dan Powell, creative technical manager at The New York Times
• Elisheba Ittoop, sound designer and composer at The New York Times

Additional Reading: 

• The Hard Fork Megamix Youtube Playlist

We want to hear from you. Email us at hardfork@nytimes.com. Find “Hard Fork” on YouTube and TikTok.

This week, the companies building artificial intelligence are facing a limit to what training data is publicly available on the internet. Will that stop them from building God? Then, a new bipartisan national privacy law proposal just dropped. We ask what’s in it. And finally, ByteDance is building new apps instead of fighting Congress’s TikTok ban.

Today’s Guests:

Trevor Hughes, president and C.E.O. of the International Association of Privacy Professionals

Additional Reading:

giants-harvest-data-artificial-intelligence.html">How Tech Giants Cut Corners to Harvest Data for A.I.

For Data-Guzzling A.I. Companies, the Internet Is Too Small

Lawmakers unveil sprawling plan to expand online privacy protections

marketing-blitz-ban.html">TikTok Turns to Nuns, Veterans and Ranchers in Marketing Blitz

We want to hear from you. Email us at hardfork@nytimes.com.Find “Hard Fork” on YouTube and TikTok.

This week we look at how AI is affecting jobs. As companies start announcing AI-related job cuts and experimenting with customer service bots, economists are placing bets on whether AI will lead to major gains for companies and workers. Some are even predicting it will help rebuild the middle class.  Then, multidisciplinary artist and filmmaker Paul Trillo joins to talk to us about his experience as part of a select group of testers granted early access to Sora, Open AI’s video generation tool. And finally, Kevin explains what happened when a Microsoft developer stumbled on a huge cyber security breach.

Today’s Guests: 

• Paul Trillo, multidisciplinary artist, writer and director 

Additional Reading: 

• tech-economy.html">How One Tech Skeptic Decided A.I. Might Benefit the Middle Class
• intelligence-productivity.html">Will A.I. Boost Productivity? Companies Sure Hope So.
• Paul Trillo’s Sora Video, The Golden Record
• cyberattack-linux.html">Did One Guy Just Stop a Huge Cyberattack?

We want to hear from you. Email us at hardfork@nytimes.com.Find “Hard Fork” on YouTube and TikTok.

Warning: The second segment of this episode includes mentions of suicide. If you are in crisis please call the suicide and crisis lifeline at 988 or you can contact the Crisis Text Line by texting TALK to 741741.

This week, we look at a mess of corporate drama in artificial intelligence. Stability AI has announced that its founder and C.E.O., Emad Mostaque, is leaving the company. Meanwhile, Microsoft hired away two of the co-founders and much of the staff of Inflection, without actually acquiring the company itself. Both moves surprised tech insiders. Then, we talked with listeners who had something to say about our interview with Jonathan Haidt on smartphones, social media and young people. And finally, we examine the true motives behind “Shrimp Jesus” and other hugely popular images on social media that were generated with artificial intelligence.

Today’s guests:

• Jordan Lucero, a high school student
• Maya Rayle, a graduate student
• Jack Campbell, a college student
• Brendan Kelley, a high school digital coach

Additional Reading: 

• The indie AI companies are falling apart
• How Spammers, Scammers and Creators Leverage AI-Generated Images on Facebook for Audience Growth

We want to hear from you. Email us at hardfork@nytimes.com. Find “Hard Fork” on YouTube and TikTok.