|
We're back from break. What did we miss? On the podcast this week: An open source piece of software that is saving the internet from AI bot scrapers, and a new facial recognition app ICE is using. In the bonus section for supporters, we do a lightning round of a bunch of our recent stories about chatbots and how to trick them. Listen to the weekly podcast on Apple Podcasts, Spotify, or YouTube.
Contact us here about advertising in the free version of our newsletter.
Los Angeles readers, we have a very important announcement: Come hang out with us IRL on July 30! Become a paid subscriber at the supporter level and reply to this email for a code to use at checkout. Otherwise, you can grab a regular admission ticket for entry. THE 'ALL CLAIMS DATABASE'ICE agents have access to a massive database of health and car insurance claims and are using it to track down people they want to deport, according to internal ICE material viewed by 404 Media. The database, which contains details on more than 1.8 billion insurance claims and 58 million medical bills and growing, includes peoples’ names, addresses, telephone and tax identification numbers, license plates, and other sensitive personal information.  Screenshot via Anubis BE GONE BOTSXe Iaso decided to develop Anubis, a “program is designed to help protect the small internet from the endless storm of requests that flood in from AI companies,” after discovering that her own Git server was struggling with AI scrapers, bots that crawl the web hoovering up anything that can be used for the training data that power AI models. Like many libraries, archives, and other small organizations, Iaso discovered her Git server was getting slammed only when it stopped working.  A sign at Wilson's Creek, a Civil War battlefield in Missouri. SlapTheWorld, CC0, via Wikimedia Commons HELP PRESERVE HISTORY The Trump administration is now targeting National Parks signs that educate visitors about anything other than “beauty” and “grandeur,” and demanding they remove signs that mention “negative” aspects of American history. Volunteer preservationists launched Save Our Signs, a project that asks park visitors to upload photos of placards, signage and monuments on public lands—including at national parks, historic sites, monuments, memorials, battlefields, trails, and lakeshores—to help preserve them if they’re removed from public view.  Amina, an AI refugee constructed by a class at United Nations University. 404 Media screengrab. CAN'T MAKE THIS UPThe United Nations University Center for Policy Research created a pair of AI avatars in an attempt to help people learn more about what refugees are facing. The UN itself says a major problem facing refugees is their dehumanization for political gain or convenience. The group also tested a soldier persona called “Abdalla,” which “simulates the behavior and decision-making patterns of an actual combatant, offering negotiators and mediators a possible tool to train for future high-stakes negotiations.” READ MORE404 IN THE WILDJason recently went to the Publishing Insider Summit and talked about newsletter trends (so meta!) alongside our Aftermath pal Nathan Grayson. And Joe went on the Marketplace Tech podcast to discuss how data analytics company Palantir has been working with ICE to track migrants in real time, and was on the What Next podcast to talk about how airlines sell your data. Replying to Researchers Jailbreak AI by Flooding It With Bullshit Jargon, CJ Tweedie wrote: “I like how this is essentially breaking down an LLM’s defences with the sort of waffling nonsense and fake references often spat out by an LLM. But it also shows the versatility and dynamism inherent in language itself.”
And in response to The UN Made AI-Generated Refugees, RM wrote: “UNU-CPR discovers an innovative and efficient way to quickly and at global scale make people scream ‘Just talk to people! Just fucking talk to a real person!’ at their computers.”
Just fucking talking to real people: That’s the 404 Media Promise. BEHIND THE BLOGThis is Behind the Blog, where we share our behind-the-scenes thoughts about how a few of our top stories of the week came together. This week, we discuss choosing what to cover, fishing expeditions, and the library. JOSEPH: There’s an interesting cybersecurity story going on: a politically-motivated, right wing extremist hacker broke into Columbia University, stole swathes of applicant and other data, and then leaked parts of it. That’s how you got this New York Times article about New York mayoral hopeful Zohran Mamdani applying to the university as Asian and African American. Predictably, some people are reading that in bad faith, when in reality it shows how stupid and regimented application processes can be in a diverse society. This week The Verge published a piece called The Columbia hack is a much bigger deal than Mamdani’s college application, which basically elaborates on that point. Read the rest of Joseph's Behind the Blog, as well as Sam, Emanuel and Jason's, by becoming a paid subscriber.
|
