Cover image via Follow us on Instagram, TikTok, and WhatsApp for the latest stories and breaking news.
A top safety researcher at Anthropic has estimated a greater than 10% chance that artificial intelligence could eradicate humanity within a decade
Evan Hubinger shared his grim forecast on X, noting that while current models pose a low risk, the technology's rapid self-improvement could soon become an existential threat.
The warning arrived alongside the high-profile departure of Jacob Coxon, another researcher who recently left Anthropic.
"Neither company is acting responsibly," Coxon wrote, referring to his former employers Anthropic and OpenAI.
He warned that superhuman systems could soon hack global infrastructure, revolutionise entire fields, and amass immense power
Dame Wendy Hall, a computer scientist advising the UN, expressed shock at the social media posts. However, Hall suggested they might be a cynical PR stunt ahead of highly anticipated stock market debuts.
"Why would someone want to say that? I would plead with investors not to invest in this company if that is their value system," she told the BBC.
The warnings prompted urgent calls for government intervention and a multinational treaty on superintelligence
Darren Jones, a former UK chief secretary, recently petitioned UK Prime Minister Andy Burnham to establish global safety guardrails before the technology outpaces regulation.
This came amid reports that Anthropic intentionally withheld its latest model from the UK's AI Security Institute (AISI) for safety assessment.
With OpenAI, Meta, and Anthropic all disclosing recent hacks carried out by autonomous AI tools, the tech industry's grip on its own creations appeared increasingly fragile.
