An AI startup spent about $50,000 to run a one-day, 24-hour news experiment on X with AI anchors reading Reuters stories, and it misidentified a woman in a Reuters clip as President Trump. Meanwhile, a real channel of AI-generated shows and AI-generated ads quietly launched on Roku and runs around the clock. And The New York Times reports the tech giants that built these tools are now racing to clean up the slop: Spotify removed 75 million tracks, LinkedIn built a report button, TikTok and Pinterest added dials to turn AI content down.

We ask the obvious question: who is this for? And what happens to trust when you can't tell whether a camera, or a person, was ever there.

Tool of the Week: Claude connectors. Claude can plug into other apps and tools and do the work from one chat, including a viral workflow that turns a line drawn on a map into drone-style video.

Making AI make sense, one conversation at a time.

A company with 10,000 paying customers, on track for $10 million this year, and zero employees. The Wall Street Journal says solo founders like this are multiplying, and this week Mark Zuckerberg published Meta's vision for the AI future predicting exactly that: smaller companies, more of them.

In this episode: how college students are actually using AI (and the rules they're begging for), the rise of the one-person company, the Michigan town suing an AI data center over noise that never stops, plus what Meta's "personal superintelligence" pitch means for regular people.

Hit "share" on an AI chat lately? Hundreds of those conversations just turned up on Google, including private resumes and work projects. Olivia and Dave break down what happened and how to protect yourself, plus LinkedIn's new "AI slop" report button and Mark Zuckerberg's rare media blitz arguing the AI industry is too gloomy. Then: poet and playwright Hannah Silva on writing her acclaimed memoir My Child, the Algorithm with an AI before ChatGPT existed, and what a machine taught her about love, motherhood and being human.

This week, poet Joseph Fasano joins us to read the poem the internet keeps passing around: "For a Student Who Used AI to Write a Paper." We talk about what students are really reaching for when they reach for a shortcut, and what poetry can do that a chatbot never can.

Plus: China cracks down on AI "boyfriends" and "girlfriends," and the real reason is the birthrate. The data center boom is minting millionaires in rural Pennsylvania while spiking water bills next door. And Anthropic bets its new ad campaign on your doubts about AI.

As always: no hype, no doom, just AI explained for real people.

Guest: Joseph Fasano, poet, novelist and songwriter. Find his poem at poets.org and his books wherever books are sold.

This week on Prompt Response, we dig into the AI cheating scandal at Brown University, where a take-home midterm averaged 96 and the in-person final collapsed to 48. What happened in between says everything about AI and education right now. Plus: Meta pulled its brand-new Muse Image feature after just three days of backlash (we'll show you the Instagram setting to check right now), and The Washington Post's Michael J. Coren on "cognitive surrender" - and how to use AI without letting it think for you.

Millions of young people are telling AI chatbots how they feel before they tell a parent, a friend or a doctor. Yale psychologist Marc Brackett, founding director of the Yale Center for Emotional Intelligence and author of "Dealing with Feeling" and "Permission to Feel," says that's not a technology story - it's a culture story, and it's on us. We talk about why kids trust the bot and how to actually get real help when the system feels overwhelming.

Plus: a man's life is saved thanks to AI, Google reimagines the Declaration of Independence as a group project and a World Cup moment you have to see AI recreate.

Making AI make sense, one conversation at a time.

The government is telling AI companies to hold back their most powerful models. China just released one anyone can download for free that rivals those capabilities. And the data centers powering all of it are showing up on farmland, using local water and straining local power grids. Axios Senior AI Reporter Madison Mills joins us to connect all three stories and explain what it means when AI stops being a software story and becomes a national security and infrastructure story.

For the first time ever, nearly half of Americans now use AI chatbots, and most of them are worried about it. Brands are secretly deploying fake AI influencers to sell you things. And the CEOs of the world's biggest AI companies just got seats at the G7 next to actual heads of state. This week, we connect the dots from your phone to the world stage, and Dave teaches Olivia everything she needs to know about Huupe, the AI basketball hoop that tracks your shot like an NBA statistician.

In this episode of Prompt Response, we explore why Anthropic’s latest AI model became a national security concern, how McDonald’s is bringing AI to the drive-thru, why deepfakes are creating a growing crisis of trust online and the backlash facing Martin Scorsese over AI’s role in filmmaking. Plus, we cover NotebookLM as our tool of the week.

In our very first episode of Prompt Response, we explore how AI is rapidly expanding beyond tech and into every corner of society. We discuss Anthropic’s plans to go public, why universities are rethinking education in the age of ChatGPT, and the debut of an AI-generated film at the Tribeca Film Festival. Plus, we take a look at Claude’s new “Cowork” feature and what it means for the future of AI agents that can actually complete tasks on your behalf.