Pakistan, Sept. 9 -- Anthropic researcher Evan Hubinger has warned that artificial intelligence could wipe out humanity, placing the risk above 10% within the next decade. He made the alarming claim while responding to former colleague Jacob Coxon, who recently resigned over concerns about AI safety.

Coxon, who worked on training new AI models at Anthropic, accused AI companies of taking dangerous risks with rapidly developing technology. He said researchers were moving toward self-improving superintelligence without fully understanding its possible consequences.

He also warned that future AI systems could become more powerful than humans in several areas. According to Coxon, such systems could potentially hack computer systems, transfo...