Meta Killed Its Own Instagram AI Tool After Users Realized What It Was Doing
Two of the biggest names in tech are in legal and ethical trouble this week — and both stories trace back to the same question: who gets to decide what happens to your data? Today’s news covers a botched Instagram feature that let strangers generate fake images of you, and a lawsuit that claims OpenAI’s leadership helped steal Apple’s most sensitive secrets.
Meta Launched an AI Deepfake Tool on Instagram, Then Quietly Pulled It
Meta rolled out a feature this week called Muse — an AI image generator built directly into Instagram. The problem? It was designed to create realistic images of real people using their public Instagram photos, without asking those people first. Within days, the backlash was loud enough that Meta disabled it entirely.
Here’s what made this particularly unsettling. Muse was opt-in by default, meaning anyone with a public Instagram account was automatically enrolled. A stranger could theoretically type your name, pull from your public posts, and generate images of you doing things you never did. That’s not a hypothetical privacy risk — it’s a concrete one that affects anyone with a public profile, from local business owners to casual creators.
For everyday users, the lesson is blunt: features that use your public content to train AI systems are showing up faster than most people expect. You don’t need to be a celebrity to be affected. Anyone visible online is now a potential subject. Meta has since removed the feature, as TechCrunch and The Verge both reported, but the fact it launched at all signals where platforms are pushing.
Why this matters: This wasn’t a bug — it was a deliberate product decision that Meta reversed only after public pressure. That pressure worked, which means user awareness actually has teeth.
“Meta removes Muse Image feature that automatically opted public Instagram accounts into AI training.”
Apple Is Suing OpenAI, and the Allegations Go All the Way to the Top
Trade secrets are the confidential internal knowledge a company builds over years — think proprietary hardware designs, engineering blueprints, internal research. Apple’s lawsuit, filed in federal court in Northern California, claims OpenAI didn’t just accidentally receive this information. According to Apple, the theft was organized at senior leadership levels, involving former Apple employees who took that knowledge with them to OpenAI.
This isn’t a story about a rogue intern downloading files before leaving. Apple’s complaint, as Reuters and CNBC reported, argues the scheme ran through multiple levels of leadership. That framing matters legally because it makes the alleged misconduct harder to dismiss as an isolated lapse. It suggests coordination, not coincidence.
For people who use Apple products, the practical stakes are about what comes next. If the lawsuit has merit, it could reshape what OpenAI is allowed to build — particularly around hardware integration, which is increasingly where AI competition is heading. Apple and OpenAI are already connected through the Siri partnership announced in 2024. A bitter legal fight could complicate that relationship significantly.
Why this matters: A lawsuit this serious, aimed this high up the org chart, signals that the AI talent wars have real legal consequences now. Companies can’t just hire from competitors and assume nothing follows.
“Apple sues OpenAI alleging trade secret theft, says scheme was ‘at every level.’”
Also Happening in AI
The AI infrastructure story continued to grow this week on a massive scale. South Korean chipmaker SK Hynix — a major manufacturer of the memory chips that power AI systems — raised $26.5 billion through a US stock listing, the largest foreign company IPO in American history, per TechCrunch. The company is being urged to build new fabrication plants on US soil. Meanwhile, on a very different scale, solar company Sunrun is reportedly offering homeowners payment to install small AI computing equipment in their homes — essentially turning residential properties into micro data centers, according to The Verge. And inside OpenAI, Wired reports that Johannes Heidecke, the company’s head of safety systems, is leaving following an internal reorganization — a notable departure given the ongoing public scrutiny over how seriously AI labs take safety commitments.
What to Watch
The OpenAI safety departure and the Apple lawsuit are worth watching together. OpenAI is losing a senior safety leader at the same moment it’s facing allegations about how it acquires talent and knowledge — and those optics will matter in any coming regulatory hearings. Watch whether the EU or US Congress uses this lawsuit as a hook to push harder on AI governance legislation this fall. The Muse controversy also isn’t over: Meta will almost certainly reintroduce something similar with slightly adjusted defaults, so the real question is whether regulators move to require genuine opt-in consent before the next version launches.