In an essay published on Saturday, Dario Amodei emphasized the need for artificial intelligence companies to slow down the pace of model advancements to prioritize safety.
He outlined a three-step plan that includes granting third-party evaluators access to verify safety practices, encouraging collaboration among AI firms in democratic nations to establish safety standards, and fostering dialogue between democratic and authoritarian governments.
Amodei's call for a more deliberate approach follows a recent resignation from an Anthropic researcher, who expressed fears about the existential risks posed by AI. While Amodei acknowledged that some aspects of his plan may be challenging to implement, he believes that taking the time to align AI technologies with human values is crucial.
His views received support from notable figures in the industry, including OpenAI's Sam Altman and Elon Musk, who both recognized the importance of pacing AI development. Amodei's stance reflects a growing consensus among researchers that while AI holds the potential to significantly enhance human life, the associated risks must be managed carefully to avoid catastrophic outcomes