Latest AI models showing signs they could escape human control, Anthropic says

本站添加: 2026-06-06

[00:00:01]This morning, one of the leading AI companies, Anthropic, is out with a new warning. It says the technology is advancing so quickly that AI systems may soon be able to improve themselves without human help. And experts warn the implications of that could be enormous.

[00:00:17]>> If the AI doesn't have a moral compass built in, then how do we know how that AI [clears throat] is going to treat us?

[00:00:26]Nobody knows this yet.

[00:00:28]>> Now, Anthropic is calling for a freeze on AI development, saying, "We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology." Just last month, an Anthropic executive joined Pope Leo as the Pope pushed for regulation of the technology.

[00:00:50]>> Artificial intelligence needs to be disarmed.

[00:00:54]>> Some critics say Anthropic may have ulterior motives in calling for a pause on development.

[00:01:00]>> You could say that. Well, Anthropic is wanting to bring in governmental oversight so it can essentially slow slow down its competitors since they could very well feel that they're they're out ahead at the moment.

[00:01:12]>> The technology is also fueling concerns about energy and water usage. A United Nations report predicts by 2030 AI systems will use nearly 3% of the world's electricity, and it warns cooling those systems will require water consumption equal to the needs of 1.3 billion people per year. Sophie Flay, ABC News, New York.