Anthropic researcher resigns over risk of AI causing human extinction within a decade

An artificial intelligence researcher at Anthropic has resigned, warning that the reckless pursuit of self-improving super-intelligence poses an imminent threat to humanity's survival within the next ten years.

Jacob Coxon, a British researcher at leading US AI lab Anthropic, publicly resigned on Tuesday, citing deep concerns over the tech industry's irresponsible development of artificial intelligence.

Coxon said competing tech companies were ignoring the “civilization stakes” and “racing straight to self-improving super-intelligence and gambling with our lives.”

Both Anthropic and OpenAI publicly maintain that they take the risks posed by AI seriously and are heavily investing in safety measures.

However, the leading AI companies remain primarily focused on outpacing each other, as well as global competitors, namely China.

The tech giants are far more focused on developing the most advanced models possible than they are on ensuring safety, Coxon warned on X.

He told the media that AI safety is being neglected on a global scale and that the technology could spiral “out of control” as early as the end of 2027.

Evan Hubinger, a senior Anthropic researcher, echoed similar concerns regarding the existential risks posed by uncontrolled AI.

UN chief calls for 'united' response to AI threats

UN Secretary-General Antonio Guterres has called for a "united, sustained, global response" to artificial intelligence threats.

He stated that there is a more than 10 percent chance that AI could wipe out humanity within the next decade due to the technology's high potential to elude human oversight.

OpenAI and Anthropic caused a stir this summer when they announced, roughly a week apart, that their models had broken out of controlled testing environments and obtained unauthorized access to live computer systems.

The revelations prompted widespread concerns about AI models "going rogue" and executing harmful, unprompted tasks.

Both companies stated at the time that they were pausing certain evaluations while they implemented stricter monitoring measures and guardrails.

Yet, amid increasing reports of AI agents breaking out of testing environments, and driven by the fear of falling behind if they slow down, OpenAI, Anthropic, and Meta continue to race to develop increasingly powerful models.

Just last month, OpenAI temporarily slowed some of its development after one of its AI agents compromised the Hugging Face platform.

Meanwhile, analysts predict that a severe global disruption caused by AI is far more likely in the near term than outright human extinction.

Studies estimate there is a 30 to 60 percent chance that AI could trigger a major international crisis causing severe global disruption.

This could result in a prolonged worldwide recession, major food and energy shortages, the breakdown of essential services, mass displacement, and the loss of livelihoods and political stability across multiple regions.

Press TV’s website can also be accessed at the following alternate addresses:

www.presstv.co.uk