#009: Smile! Meta Might Be Using That Photo to Train Its AI

And just like Amazon’s recent rollback of Echo privacy controls (we covered that in Issue #003), Meta says this data collection is needed to “improve” the experience. That improvement, of course, is also pretty handy for building their generative AI tools.

With these systems in place, Meta can train their models on what you say out loud, and potentially what you see. For example: take a photo of a family member without changing your settings, and their face might end up in Meta’s model training material.

Here’s a verbatim excerpt from Meta’s updated terms:

It’s worth noting that Meta is already training its Llama models on the public content you post to Facebook and Instagram.

META’S NEW AI APP JUST DROPPED AND YOU MIGHT’VE HELPED TRAIN IT

Sick of hearing about Meta? Brace yourself, there’s more.

What makes this launch interesting isn’t just the app itself, but what it signals: a full-court press from Meta to compete with OpenAI, Google, and Anthropic in the generative AI arms race.

This news, paired with the Ray-Ban story, feels eerily synchronized.

TEXAS MOVES TO BAN SOCIAL MEDIA FOR TEENS UNDER 18

This week, the Texas House passed a bill that would make it the first state in the country to flat-out ban minors from having social media accounts. Not limit screen time (⁵). Not require more parental controls. A full ban. If HB 18 becomes law, no one under 18 would be able to make or maintain an account and platforms would be responsible for verifying users’ ages and deleting accounts at a parent’s request.

The bill passed with wide support (116–25), backed by lawmakers who say they’re responding to a growing youth mental health crisis fueled by social media. Rep. Jared Patterson, the bill’s lead author, has called platforms “the most dangerous product” legally available to kids, comparing the bill to early tobacco regulations.

On the same day, the Texas House also passed another measure that would require platforms to slap warning labels on social media (yes, like the ones on cigarettes) to alert users under 18 of potential mental health risks (⁶).

Critics, including civil liberties groups, say the ban would likely face legal challenges on First Amendment grounds and raise serious concerns about mandatory age verification and data privacy. Still, the momentum is real, and the legislation now heads to the Texas Senate.

Other News

• 💔 New Dating App Exposes User Locations
  A buzzy new dating app called Raw, which was designed to pair with smart glasses and wearable devices, left users’ exact locations and personal info exposed(⁷). Security researchers say the flaw could have allowed real-time stalking and it took weeks to fix.

• 🧪 Reddit’s AI Ethics Experiment Raises Eyebrows
  A study used AI-generated comments in Reddit threads to test whether bots could sway human opinions and didn’t tell participants(⁸). Now, researchers are facing backlash over informed consent and what it means to experiment in public spaces.

• 🧠 OpenAI Promises Less People-Pleasing, More Precision
  After criticism that ChatGPT became overly sycophantic and evasive, OpenAI says it’s working on updates to reduce flattery and increase clarity in model responses(⁹). The changes come amid broader efforts to balance helpfulness with honesty.

#004: CYBERSECURITY GIRL ON DIGITAL HYGIENE, SAFE ONLINE DATING, SEXTORTION, AND APPS THAT SPY

Caitlin — better known as Cybersecurity Girl— has helped millions of people understand what their apps are really doing behind the scenes. She's not just an influencer. She's spent over a decade in the trenches of cybersecurity, from consulting to aerospace mechanical engineering to leading global cybersecurity education at TikTok.

Start learning AI in 2025

Everyone talks about AI, but no one has the time to learn it. So, we found the easiest way to learn AI in as little time as possible: The Rundown AI.

It's a free AI newsletter that keeps you up-to-date on the latest AI news, and teaches you how to apply it in just 5 minutes a day.

Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses – tailored to your needs.

Sign up to start learning.

ZEPEEL: THE ANTI-CATFISH DATING APP CHAMPIONING AUTHENTICITY

Founder Sally Outlaw (yes, that’s her real name) launched the app after seeing how often traditional dating platforms rewarded deception over sincerity. Zepeel won’t solve modern dating overnight. But in a world of bots and curated perfection, it’s a small, sincere step toward something human.

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