The Download: how people really use AI, and Flock’s design choices

This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology. AI companies like Anthropic and OpenAI regularly publish reports on how people are using their products. But they only release the data they want us to see, AI researchers say, and there’s no independent source to corroborate it.

A new research project called the AI Observatory aims to fill in the gap. Its analysis shows many more sensitive behaviors than are captured in reports from major AI companies, which focus more on work than on personal use. The researchers also found significant differences between models.

People were more likely to turn to Anthropic for coding, Gemini for social and roleplay uses, and ChatGPT for homework assistance. Here’s what the AI Observatory reveals about how people use AI. Flock Safety, the police-tech giant known for its network of some 120,000 automatic license plate readers around the US, recently announced changes to its platform.

The updates are meant to prevent officers from using it for illegal or illegitimate purposes, including stalking. Amid all this, there have recently been several arguments defending Flock: If these cameras help solve crime, is it as big a deal? On a good day they might help catch a kidnapper, and if not, they’re simply snapping pictures of my car that nobody will bother to look at.

But this all skips over a more important question: What kind of crime-fighting system has Flock chosen to build? Its network works the way it does because of decisions about what information to collect, who can search it, how long to keep it, and how widely to share it. Those decisions set the terms of the bargain between security and civil liberties.

Find out what a narrower Flock system could look like. This story is from The Algorithm, our weekly AI newsletter. . I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology. 1 A child privacy trial starting today could change Meta foreverMore than half of the states in the US have joined the lawsuit. (CNBC)+ They say Meta deliberately designed addictive social networks. (Guardian)+ And want changes including ending "like" counts and infinite scroll. (BBC)+ Four of them are demanding $1.4 trillion in damages. (WP $)2 Nvidia has committed up to $105 billion to OpenAI’s Ohio data centerIt's slated to cost up to $500 billion and come online in 2028. (CNBC)+ OpenAI will lease the eight-gigawatt site for 20 years. (NYT $)+ How virtual power plants could provide energy for data centers. (MIT Technology Review) 3 Tesla is set to launch Cybercab robotaxi rides in Austin this monthThe rollout could begin with employee rides on public roads. (Information $)+ The EVs will then enter Tesla’s robotaxi service ‌a few days later. ( $)+ The Tesla Semi could also be a big deal for EVs. (MIT Technology Review) 4 Unitree has unveiled a humanoid it claims is faster than any humanThe new “Superman” robot can cover 12.66 meters in a second. (Gizmodo)+ It arrives ahead of the Chinese company’s IPO on Wednesday. ( $)+ US robot startups are struggling with new China restrictions. (Rest of World) 5 A tracked rare book shipment led to Amazon’s AI training operationAmazon scans and destroys books at the facility to train AI. (404 Media $)+ Rare books provide unique, high-quality training data. (Ars Technica)+ But feeding them to AI risks creating a cultural void. (New Scientist $) 6 China wants its data to shape what the world’s AI knowsBeijing is distributing datasets reflecting “mainstream Chinese values.” (NYT $)+ What’s next for Chinese open-source AI? (MIT Technology Review) 7 An AI studio wants to become the HBO of adult contentRogue’s AI video tool produces uncensored adult content. (Wired $)+ AI is creating new risks for porn actors. (MIT Technology Review)8 Women are being left behind in the AI jobs boomThey accounted for just 26% of new AI hires last year. (Axios)9 AI has revealed new clues to how breast cancer progressesThe findings could help predict how tumours will develop. (Independent) 10 An unearthed video suggests Apple is adding cameras to AirPodsIt shows AirPods identifying a book using visual intelligence. (Gizmodo) —Writer Carmen Hermosillo made a prescient observation about life online in a 1994 essay, quoted by the Guardian in a story on warnings we ignored about the digital age.