Did OpenAI give up on safety?

Careful: The contents of this email are HOT! 🔥

(We earn when you visit featured links.)

[Sponsor FryAI]

This link leads to one of the most interesting things I’ve seen in AI recently.

What’s cookin’? OpenAI has reportedly disbanded its “preparedness” team, the group responsible for studying whether its most advanced AI models could create serious risks and finding ways to reduce those risks. According to the report, that work will now be split across existing teams focused on specific areas, such as cybersecurity and biological risks. The move comes during a period of major change at OpenAI as the company moves closer to a potential IPO. Several safety-focused groups and leaders have left or been reorganized in recent years, including teams focused on AGI readiness and superalignment.

🤔 Hunter’s take: The concern here is not that safety work is disappearing overnight. It is that safety may become easier to dilute when it is spread across product teams under pressure to ship. When safety becomes everyone’s responsibility, it becomes nobody’s responsibility. The more powerful AI gets, the more independent safety oversight matters. Otherwise, the people building the engine are also checking the brakes.

What’s cookin’? Hollywood’s main trade group has made a new agreement with ByteDance, the company behind TikTok, to help stop AI tools from copying movies, characters, and actors without permission. The deal covers ByteDance products like TikTok, CapCut, Dreamina, and its AI video tool Seedance. The new agreement is supposed to add stronger safety rules, or “guardrails,” so people cannot easily use ByteDance’s tools to make fake clips using famous actors, movie characters, or protected Hollywood content.

🤔 Hunter’s take: AI video is moving from silly internet clips to something that can look surprisingly real. That creates a big question: who gets to control a person’s face, voice, or character? For Hollywood, this is more about money. But for everyone else, it is also about consent. Just because AI can copy someone does not mean it should.

Hi — I’m Hunter, a PhD candidate whose work has appeared in major academic journals and popular tech outlets. I founded FryAI to make staying ahead of AI clear, accessible, and fun.

How close are we to the point where AI surpasses human intelligence across nearly every domain and begins advancing beyond our control or prediction?