Human reviewers need support to keep AI models from going rogue

Bloomberg reports that AI lab staff require backing from the top to provide successful oversight of the technology, highlighting the challenges of human review in AI safety.