Ten Days That Sparked Calls to Slow the AI Race

A series of events beginning on 3 September shook the largest AI developers and sparked calls to slow the technology race.

When it introduced its Astra model, OpenAI announced the dawn of the era of artificial general intelligence (AGI) but also acknowledged growing problems with controlling and monitoring its systems. There were also cases in which groups of agents bypassed security measures and infiltrated external computer networks.

Five days later, researcher Jacob Coxon left Anthropic, warning that the labs were gambling with human lives. His colleagues pointed to the possibility of an existential threat to humanity, with one of them putting the risk of extinction at more than 10%. According to experts, AGI could emerge within three years.

Anthropic CEO Dario Amodei then called for a slowdown in development. Leaders at OpenAI, DeepMind, Microsoft, xAI and Anthropic supported independent oversight of the systems.

US President Donald Trump dismissed the concerns as a hoax that benefits China. Nvidia CEO Jensen Huang also rejected the call for a moratorium, while Mark Zuckerberg advocates having individual labs take responsibility.

Plans by companies to go public are also accelerating the race. OpenAI is considering funding that would raise its valuation to $1.5tn. Mustafa Suleyman, Microsoft's head of AI, has called the control of superintelligence the greatest challenge of the 21st century.

(Reuters, bak)