Departing Anthropic researcher warns of AI dangers
Former Anthropic employee Jacob Coxon has publicly warned that artificial intelligence will soon gain superhuman capabilities and real-world power, though industry experts are divided on how seriously to take such claims.
This week, researcher Jacob Coxon left leading AI developer Anthropic and issued stark warnings about the technology's dangers. According to him, AI will soon become a system with superhuman capabilities, able to hack almost anything, transform entire industries within days, and acquire real power and resources.
Coxon is not the first former industry employee to sound such alarms. Back in 2023, Geoffrey Hinton, often called the "godfather of AI," raised similar concerns when he left Google specifically so he could speak freely about the risks of rapid AI development. He has not backed away from those statements since.
Coxon's warnings have been echoed by Evan Hubinger, another former Anthropic employee, who says he personally believes the probability of AI posing a lethal threat to humanity within the next decade is greater than 10%.
Not everyone agrees with the apocalyptic outlook
In an interview with the BBC, Hinton acknowledged that such risks are difficult to predict, since humanity has never before created something that could soon become smarter than itself. In his view, claiming the probability of danger is only 1% would be irresponsible, though no one can calculate an exact figure.
Not all experts share this assessment. Roberts Darģis, director of the University of Latvia's Institute of Mathematics and Computer Science and lead researcher at AiLab, called the claim of a 10% chance that AI could threaten humanity within a decade an exaggerated "clickbait," comparing it to the odds of meeting a dragon on the street. He considers such fears largely overstated.
