OpenAI just leaped forward in the AI race through its partnership with Apple, announced at this week’s WWDC24 conference. Apple will install ChatGPT in its operating systems and Siri, impacting more than 2 billion active devices worldwide. This extends the reach of OpenAI, which already had approximately 180 million active users, not to mention the millions who use Microsoft’s GPT-4-powered Copilot. The announcement comes soon after a recent exodus of two senior OpenAI executives focused on AI safety and an open letter signed by 11 of OpenAI’s current and former employees asking for more protection when speaking out about AI safety concerns. These prior incidents, including the previous departure of senior executives and questions around Altman’s firing (and rehiring), suggest that the company may be prioritizing growth over AI safety. As a leader in the AI race, OpenAI’s approach to safety and security should be closely scrutinized and potentially overhauled. On May 14, co-founder and former Chief Scientist, Ilya Sutskever, and former Head of Alignment, Jan Leike, announced that they were leaving OpenAI. Sutskever and Leike were co-leading OpenAI’s Superalignment team, which seeks to not only align AI behaviors and objectives with human values and objectives, but also prevents superintelligent AI from “going rogue.” Superalignment is becoming increasingly important as artificial general intelligence (AGI) becomes an ever-increasing possibility. Their concurrent departures may be signaling that OpenAI may not be serious about safety. Leike posted on X that “safety culture and processes have taken a backseat to shiny products,” and added in another X post: “I have been disagreeing with OpenAI leadership about the company’s core priorities for quite some time, until we finally reached a breaking point.”

Full report : How Apple and OpenAI deal could actually harm AI safety and privacy in the long run.