Summary
Wired reports that OpenAI’s Superalignment team, first launched in July 2023 to prevent superhuman AI systems from going rogue, is no more. Research on the risks associated with more powerful AI models will now be led by OpenAI cofounder John Schulman.