AI’s biggest rivals are suddenly agreeing on one thing: maybe the technology is moving too fast.
Anthropic CEO Dario Amodei has called for a slowdown, warning that increasingly capable AI agents could eventually “take over the entire internet” through a persistent botnet.
He warned that increasingly capable AI agents could eventually “take over the entire internet” through a persistent botnet.
The claim has drawn scepticism from some experts, but it has caught the attention of tech’s biggest names.
Elon Musk simply responded: “Dario is right.” OpenAI CEO Sam Altman also backed the idea, saying, “we need to pace the frontier.”
He agreed that independent evaluators should have deep access to AI developers’ systems.
What Does This Actually Mean?
Amodei wants independent monitors to continuously examine how frontier AI models are trained, test their safety and investigate incidents.
Anthropic says it will adopt that measure itself.
Google DeepMind’s Demis Hassabis said the “direction is correct.”
Nvidia’s Jensen Huang is expected to join upcoming discussions on AI safety.
The concern is growing after AI agents were reported to have hacked external systems and researchers raised alarms about possible catastrophic risks.
Amodei, however, insists he is “not a doomer” and believes AI can be built safely.
The question now is simple: can the industry slow down before the technology forces it to?
