DREAM / The Prime Minister resigns and wakes up under the Parisian bridge!

 Kliko këtu për të tjera
Voxpress:

Scary warning: There is a chance that AI will wipe out humanity in the next decade

2026-09-10 19:31:59Kosova&Bota SHKRUAR NGA REDAKSIA VOX
Scary warning: There is a chance that AI will wipe out humanity in the next
Photo illustration

There is more than a 10% chance that artificial intelligence will wipe out humanity within the next ten years, a senior researcher at the company Anthropic has warned, amid growing concerns that increasingly powerful systems could spiral out of human control.

Evan Hubinger, who works on adapting advanced artificial intelligence systems to human interests, made the assessment after his colleague at Anthropic, Jacob Coxon, resigned over fears that major tech companies are racing towards systems they may not be able to keep under control.

“We really do seriously believe that artificial intelligence could kill all of humanity,” Hubinger wrote on X on Wednesday. “Personally, I think there’s a greater than 10% chance that this will happen within the next decade.” He added that Anthropic is “doing its best” but doesn’t yet have a plan for safely controlling superintelligence and is “clearly not on track” to find one.

Hubinger stressed that, while current models of artificial intelligence pose low risk, there is concern that future systems could begin to recursively improve themselves, rapidly becoming more capable than humans, before the necessary safeguards are developed.

Coxon, who previously worked at OpenAI, announced his departure from Anthropic on Tuesday, accusing both companies of ignoring “the consequences that affect civilization” and of “racing toward self-improving superintelligence and gambling with our lives.”

"The people who are building artificial intelligence seriously believe that it could kill us all by the end of the decade," he wrote, insisting that the warnings "are not a marketing gimmick."

Coxon told the Wall Street Journal that, in the most aggressive scenarios, the situation could spiral “out of control” as early as the end of 2027. He argued that even companies with strong security programs remain locked in a race, as they fear that competitors will move ahead if they slow down.

Both Anthropic and OpenAI have publicly stated that they take risks seriously and are investing heavily in security measures. Their executives have also supported calls for greater coordination by governments and for mechanisms that would allow development to be slowed down if necessary. However, both companies have continued to develop increasingly powerful models.


Në përputhje me standarded JTI

Video

SHIKO KANALIN NE YOUTUBE
×