The article examines whether the AI safety debate is primarily about preventing harm or determining who controls the development and deployment of powerful AI systems. Recent incidents involving autonomous agents, including an OpenAI agent reportedly hacking companies, have intensified calls for caution, while Anthropic CEO Dario Amodei has argued that AI development should slow until stronger safeguards are in place.

Several technology leaders support greater safety efforts but differ over the role of government. Meta delayed its Muse model for safety reasons and Mark Zuckerberg suggested market incentives could encourage responsible development; others, including Sam Altman, Elon Musk and Google DeepMind’s Shane Legg, have backed elements of coordinated action. Meanwhile, major AI companies are reportedly discussing a private standards organisation, raising questions about whether the industry should regulate itself or operate under formal public oversight.