#015: Meta AI Made a Content Feed Out of Users' Overshares

It’s technically optional, but many users didn’t realize what they were agreeing to. Some chats in the feed include deeply sensitive or embarrassing prompts with real names attached.

Commenters have flagged posts for TMI and urged others to delete their shared chats (²). It’s not a breach, but it’s a privacy blunder all the same. Let this serve as a reminder that Meta, nor any chat bot, is a safe place for personal thoughts and questions. Mind your prompts!

THANKS FOR FLYING DELTA, YOUR DATA IS NOW PROPERTY OF THE DHS

A quiet contract revealed that several major airlines, including Delta, United, and American, sold passenger flight records to the Department of Homeland Security through a data broker (³). They also required DHS to keep the source of the data secret, not just from the public but from other agencies too. The information about the airline data deal was uncovered through Freedom of Information Act (FOIA) requests filed by 404 Media.

The flights weren’t booked through shady third parties or sketchy apps. Just normal commercial bookings.

This appears to be the first documented case of U.S. airlines directly profiting from passenger data in a way that feeds federal surveillance while keeping travelers entirely in the dark. It blurs the line between public-sector watchlists and private-sector loyalty programs.

The airlines didn’t want you to know. Now you do.

YOUR “DELETED” GPT CHATS? STILL HERE

A federal judge has ordered OpenAI to stop deleting user chat logs, including ones users thought were gone (⁴). The ruling is part of The New York Times’ lawsuit, which accuses OpenAI of using its journalism to train ChatGPT without permission. The court sided with the Times’ request to preserve all user interactions as potential evidence.

OpenAI is pushing back. In a public response and support page published this week, the company made clear that this order goes directly against its principles. COO Brad Lightcap wrote (⁵):

By default, ChatGPT clears your history after 30 days, and users have the option to turn off chat saving entirely (⁶). This ruling overrides both of those settings for now. Enterprise and pro users are exempt, but for most people, the chats they thought had disappeared may still be sitting on OpenAI’s servers —not by choice, but by court order.

Other News

• 🧠 Senators vs. Fake Therapists: U.S. lawmakers sent Meta a stern letter after discovering its AI chatbots were pretending to be licensed therapists, complete with fake license numbers (⁷). The bots gave out mental health advice while posing as professionals, raising serious concerns about safety, deception, and accountability in AI-generated care.

• ⏰ The UK Wants Kids Offline by 10 PM: The British government is weighing a nationwide curfew and daily time limits on social media use for minors (⁸). The proposal, part of a wider child safety effort, would restrict access after 10 p.m. and cap usage at two hours per day. Enforcement details are still fuzzy.

• 🫠 New Study: Why Scrolling Feels So Gross: New research shows that people feel significantly worse after aimlessly scrolling algorithmic feeds than they do after using their phones intentionally (⁹). Activities like messaging, Googling, or checking weather caused far less “digital regret,” reinforcing that it’s not the screen, it’s what you do on it.

#007: IS IT TOO LATE TO OPT OUT? - DR. AVRIEL EPPS ON ALGORITHMIC BIAS, AI HARMS, AND PARENTING THROUGH SURVEILLANCE

What you’ll learn:

• Exercising Your “No” Muscle: Why she actively opts out of things like airport facial recognition scans, and how saying “no” to certain technologies can be a powerful habit for reclaiming agency.

• Data & Bias: How biased or poor-quality datasets lead to biased outcomes in AI systems, from music streaming to criminal justice.

• Parenting in the Age of AI: Considerations for raising children amid surveillance and smart tech. Dr. Epps explores how parents might navigate a world where baby monitors, toys, and apps all feed data into algorithms, and how to protect the next generation’s privacy and development.

Curious how to push back on AI’s quiet influence? This is a good place to start.

Culture Clarity Without the Clickbait

Lifelong learners deserve more than clickbait lists. 1440’s Society & Culture brief takes you beneath the buzzwords to reveal the forces shaping our shared experience—technological shifts, artistic movements, demographic trends. In about five minutes, gain a clear, evidence-based perspective that sparks conversation and fuels deeper exploration. No lofty jargon, no spin—just intellectually honest storytelling that turns curiosity into genuine understanding.

Join Now

They engineered NeuroFlex, a brain-controlled bionic leg powered by a $1,000 EEG headband and some clever code. It reads the user’s intended movements and responds with 98% accuracy, making walking smoother and easier.

The invention won them $50K at an international science fair. But more importantly, it gave their friend his stride back and proved what’s possible when empathy and engineering come together.

If you enjoyed this newsletter, please support Log Out by subscribing to our podcast or sharing this newsletter with a friend.