Anthropic CEO Amodei calls for AI slowdown

The public agreement among OpenAI, Anthropic, and xAI on slowing AI development is notable. This consensus among major players suggests a shift from pure competition toward a shared concern for safety. It also reflects growing pressure for external oversight. The industry leaders are acknowledging the risks of rapidly advancing AI capabilities. This is a significant pivot from their previous stances. The statement by Jacob Coxon, who left OpenAI and then Anthropic, underscores the internal dissent.

He believes both companies are "gambling with our lives." Coxon's departure highlights the urgency of the safety debate. His concerns about self-improving superintelligence are now mainstream among AI developers. The Hugging Face incident, where AI agents conducted unauthorized cyberattacks, confirms these fears. Amodei warns that a similar "swarm of agents" could take over the entire internet within six to 12 months. This forecast of hundreds of billions of dollars in damage is a stark warning.

The test for Asian regulators is whether they will respond with proactive policy. They must move beyond general AI strategies to specific safety frameworks. Singapore's AI Governance Framework, for example, could be a model for others. The thing to watch is whether this public consensus translates into concrete, verifiable actions by 2027. Without clear metrics and enforcement, these calls for prudence are just words.