More than a 10 per cent chance of humanity being wiped out by AI? That is not the prediction of an anti-technology activist, but the personal estimate of a senior AI safety researcher at Anthropic ,and it has reignited one of the most controversial debates surrounding artificial intelligence.
A senior Anthropic AI safety researcher has warned that increasingly powerful artificial intelligence could pose an existential threat to humanity within the next decade.
Evan Hubinger, Anthropic’s Alignment Science Lead, said he personally believes there is a greater than 10 per cent chance that AI could eventually kill all humans within the next decade.
Hubinger’s comments came after fellow Anthropic researcher Jacob Coxon announced his resignation and criticised the industry’s race towards self-improving AI.
Hubinger said the greatest concern was not necessarily today’s AI systems but future systems capable of recursively improving themselves at an increasingly rapid pace.
He also acknowledged that Anthropic does not yet have a complete solution for ensuring that future superintelligent systems remain aligned with human interests.
The warning has intensified debate over the so ,called AI alignment problem , the challenge of ensuring that increasingly powerful AI systems continue to operate according to human goals and safety requirements.
However, the claim remains a risk assessment rather than a prediction that humanity will definitely be destroyed. Other experts continue to disagree about how likely an AI catastrophe is and how soon it could occur.
The controversy nevertheless highlights an uncomfortable contradiction within the AI industry, the same technology being promoted as a tool for solving humanity’s biggest problems is also being described by some of its own researchers as a potential existential risk.
