OpenAI employees oppose social media push after Sora 2 release

Several current and former OpenAI researchers have publicly commented on the company's first move into social media through the Sora app. The app functions as a TikTok-style feed full of videos, but generated by artificial intelligence, including numerous deepfakes of Sam Altman. On the X platform, the researchers expressed concerns about how this launch fits into OpenAI's nonprofit mission, which focuses on developing advanced artificial intelligence for the benefit of humanity. For example, pretraining researcher John Hallman wrote on X: "AI-based feeds are scary. I won't deny that I felt some concern when I first learned we were releasing Sora 2. That said, I think the team did the absolute best job they possibly could in designing a positive experience... We will strive to ensure that artificial intelligence helps rather than harms humanity." Another researcher, Boaz Barak, who is also a professor at Harvard, replied: "I share a similar mix of concern and excitement. Sora 2 is technically amazing, but it is too early to congratulate ourselves on avoiding the pitfalls of other social media platforms and deepfakes." Former researcher Rohan Pandey took the opportunity to promote his new startup, Periodic Labs, which consists of former researchers from the artificial intelligence lab and is working to build artificial intelligence systems for scientific discovery. He said: "If you don't want to build an endless AI TikTok slop machine, but want to develop artificial intelligence that accelerates fundamental science... join us at Periodic Labs." There were also other similar posts from people such as Will Depue and Miles Brundage.

Management's Response and Sam Altman

Sora's launch underscores a fundamental tension within OpenAI that has repeatedly resurfaced. The company is the fastest-growing consumer technology company in the world, but it also operates as a frontier artificial intelligence lab with an ambitious nonprofit charter. Some former OpenAI employees argue that the consumer business can, in principle, serve the mission: ChatGPT helps fund artificial intelligence research and disseminate the technology widely. OpenAI CEO Sam Altman confirmed this in a post on X on Wednesday, where he addressed why the company is devoting so much capital and computing power to an artificial intelligence social media app. He said: "We mostly need the capital to build artificial intelligence that can do science, and we are certainly focused on AGI (artificial general intelligence) with almost all of our research effort. It is also nice to show people cool new technologies/products along the way, make them smile, and hopefully make some money given the need for computing power." He added: "When we launched ChatGPT, there was a lot of 'who needs this and where is AGI.' The reality is nuanced when it comes to optimal trajectories for a company." In a June podcast, Altman also discussed the "great social media misalignment," saying: "One of the big mistakes of the social media era was that feed algorithms had many unintended negative consequences for society as a whole and perhaps for individual users as well. Even though they did what the user wanted—or what someone thought users wanted—at that particular moment, which was to keep them on the page longer."

Risks and Features of the Sora App

The question of when OpenAI's consumer business will outweigh its nonprofit mission remains open. In other words, when will OpenAI say no to an opportunity to make money and grow the platform because it conflicts with the mission? This question looms as regulators scrutinize OpenAI's transition to a for-profit structure, which the company needs to raise additional capital and potentially go public. California Attorney General Rob Bonta said last month that he was "particularly concerned with ensuring that OpenAI's stated nonprofit safety mission remains central" during the restructuring. Skeptics dismiss OpenAI's mission as a tool for attracting talent from major technology companies, but many OpenAI insiders insist that it is the main reason they joined the company. For now, Sora's footprint is small; the app is only a few days old. But its debut represents a significant expansion of OpenAI's consumer business and exposes the company to the incentives that have plagued social apps for decades. Unlike ChatGPT, which is optimized for usefulness, OpenAI says that Sora is built for entertainment—a place to generate and share artificial intelligence clips. The feed resembles TikTok or Instagram Reels, platforms notorious for their addictive loops. OpenAI insists that it wants to avoid these pitfalls, stating in a blog post about Sora's launch that "concerns about doomscrolling (endless scrolling), addiction, isolation, and feeds optimized for real life are top of mind." The company explicitly says that it does not optimize for time spent in the feed and instead wants to maximize creation. OpenAI promises to send reminders to users when they scroll for too long and, above all, to show them people they know. This is a stronger starting point than Meta's Vibes—another feed of short artificial intelligence videos released last week—which looks as though it was rushed out without as many safeguards. As former OpenAI policy lead Miles Brundage points out, it is possible that there will be both good and bad applications of AI video feeds, much as we have seen in the chatbot era.