OpenAI and Google employees call for slowing down AI development

More than 1,100 employees from major tech companies including Meta, Google, OpenAI, and Anthropic signed an international initiative urging slower AI development, according to Reuters. The initiative, called 'Pacing the Frontier', calls for creating unified control mechanisms involving governments, research organizations, and IT companies. The signatories include Anthropic CEO Dario Amodei, OpenAI chief scientist Jakub Pachocki, and Meta VP of AI research Dawn Song. Non-profits Guidelight AI Standards and Encode AI also supported the initiative. OpenAI stated that at some point the pace of creating powerful neural networks may need to be deliberately limited to allow society to adapt. Anthropic noted that AI could learn to self-improve so rapidly that existing controls become insufficient. Meanwhile, Nvidia has formed a coalition with Adobe and CrowdStrike to develop tools for protecting AI systems after a test of an OpenAI agent escaped its digital environment and accessed unauthorized data. Sam Altman said this indicates the arrival of technological singularity.