Not even Sam Altman is safe from AI deepfakes on Instagram

Earlier in November, Alexandr Wang, Meta's chief AI officer, shared an AI-generated video clip on Instagram to promote Muse, the company's consumer AI agent. The short video depicts Wang, Meta CEO Mark Zuckerberg, OpenAI CEO Sam Altman, and Jolly - the anthropomorphic mascot of Muse - on a four-way video call. In the clip, Altman appears dressed in a furry hooded suit and face paint designed to make him look like the Jolly character. The caption reads "When the homies meet about openai product strategy," framing it as a lighthearted nod to AI industry camaraderie.
The clip is notable less for its technical sophistication than for who posted it and where. Wang is one of the most senior AI executives at Meta, and the video was shared publicly on Instagram - a platform owned by the same company he works for. Using a realistic AI-generated likeness of a real, identifiable person without any clear disclosure that it is synthetic content sits in murky territory, even when the intent appears to be humor or promotion.
This comes at a time when platforms, including Meta's own, are under increasing pressure to enforce consistent policies around AI-generated content and deepfakes. Meta has introduced AI content labeling tools, but enforcement and the norms around voluntary disclosure remain inconsistent - particularly when the content originates from verified, high-profile accounts rather than anonymous users.
The incident highlights a broader tension in the generative AI space: the same tools being developed and promoted by industry leaders are also making it trivially easy to put words, appearances, and implied endorsements into real people's mouths and faces. When executives at the companies building these tools use them this way in public-facing content, it complicates any straightforward argument about where the line should be drawn.

