Meta’s chief AI officer just posted an AI-generated video of a rival CEO, and almost nobody flinched. Earlier in November, Alexandr Wang shared a short Instagram clip promoting Muse, Meta’s consumer AI agent. It shows Wang, Mark Zuckerberg, OpenAI CEO Sam Altman and Jolly, Muse’s furry mascot, on a four-way video call, The Verge AI reports. Altman joins the call dressed as Jolly, in a hooded suit and face paint.
The caption reads: “When the homies meet about openai product strategy.” It’s a joke about the AI agent race. It also shows how normal deepfakes of real people have become.
⚡ The Quick Read
- What happened: A top Meta executive posted an AI video of a competitor’s CEO to market a Meta product.
- Was it labeled? Yes. Meta requires AI videos that show people to carry an AI label, and Wang’s post has one.
- Meta’s defense: Spokesperson Tracy Clayton pointed to the label. He also noted that Altman has made his likeness available for AI generation, as he did when OpenAI launched its Sora app, which quickly filled up with fake Altman videos.
- OpenAI’s response: None. The company didn’t respond to requests for comment, according to The Verge AI.
- The real gap: Static AI images on Instagram don’t need any AI label at all.
📉 How Fast the Norms Moved
A few years ago, AI content showing real people drew public condemnation and, in some cases, threats of legal action. That’s mostly gone. The Verge AI describes AI content as “part of the texture of a new reality,” especially on social media.
Politics shows it clearly. In the run-up to the midterms, political groups spent millions of dollars on AI-generated attack ads. Now a senior executive at one of the world’s biggest platforms is using a rival’s face in a marketing clip. That’s a long way from where the industry stood in 2023.
Wang’s style plays a part here. He’s been running what some call “anti-PR” for Muse: hundreds of tweets, plus jabs at traditional marketers as “corposlops.” Faking a competitor’s CEO fits that playbook. It’s still notable when the person who sets AI strategy at Meta does it on Meta’s own platform.
🚨 Why It Matters: Instagram’s Labeling Loophole
The timing is awkward for Meta. Instagram has taken heat over AI content lately, and the problem goes well beyond executives trading jokes.
- The Mia Goth case: In October, an AI-generated image of the actress spread on Instagram. Her representative told The Verge it was “incomprehensible” that the platform let it circulate instead of focusing on the “protection and integrity of a public figure’s likeness.”
- No disclosure required: Because it was a static image, the Goth fake carried no AI label.
- Verified spreaders: A fan account with an Instagram-issued blue verification badge shared it.
- Sexualized fakes: Celebrities like Scarlett Johansson are regular targets. Meta has removed some impersonation accounts, but enforcement is patchy.
The pattern is hard to miss. Video gets labeled. Images don’t. A verified badge can make a fake look more trustworthy, not less.
🧭 My Take
The Altman clip isn’t the scandal. Altman has opted his likeness into AI generation, so this is consensual trolling between rivals who know each other. The trouble is what it signals. When Meta’s own AI leadership treats deepfakes of real people as routine marketing, it’s harder for the company to argue that it takes likeness protection seriously for everyone else.
Consent is the dividing line. Altman said yes. Mia Goth didn’t. Instagram’s rules currently can’t tell the difference if the fake is a still image.
🔮 What to Watch
- Label rules for images: Pressure is building on Meta to extend AI disclosure from video to static images. Expect that debate to sharpen.
- Likeness opt-in becomes the norm: Altman’s Sora move gives platforms a ready defense. Public figures may feel pushed to state clearly whether they allow AI use of their face.
- Deepfakes as marketing: If Wang’s “anti-PR” works for Muse, other AI companies will copy it. Rival CEOs in absurd AI skits could become standard launch content.
- Verification under review: Blue badges on accounts spreading unlabeled fakes undercut what verification is for.
For brands and creators, the lesson is practical. AI-generated content featuring real people now spreads with little friction, so check what you share and label what you make. The full story, including Meta’s comments, is at The Verge AI.