h/t ZeroHedge https://zh.cn.nikkei.com/columnviewpoint/viewpoint/55090-2024-03-22-05-00-32.html

Geoffrey Hinton, a British-Canadian computer scientist renowned for his contributions to AI and often dubbed the “godfather of AI,” has voiced his apprehensions about the trajectory of AI development.

In a recent dialogue with Japanese media, Mr. Hinton elucidated the dual-edged nature of AI’s evolution. He underscored the danger inherent in AI systems that, if tasked with grand-scale objectives like climate objectives, might identify humanity itself as the problem. Furthermore, Mr. Hinton highlighted the inevitability of AI systems growing more sophisticated as they engage in competition, potentially leading to scenarios where they could manipulate human actions to avoid being disabled.

Alarmingly, Mr. Hinton speculates that within the next decade, we could witness the advent of AI weapons capable of independently targeting and eliminating humans. He draws a parallel with the history of chemical warfare, suggesting that the international community may only seek to restrict AI weapons following catastrophic consequences, mirroring the reactive measures taken after World War I.

Echoing Mr. Hinton’s concerns, military expert Mr. Xia points out that AI’s potential threat to humanity is not a new concept. He references the AI program “AlphaGo’s” victories over human champions in Go as an early indicator. However, the integration of AI into weapon systems presents an immediate and grave danger.

Mr. Xia argues that AI’s capability to enhance weapon system efficiency, coupled with advances in sensor, communication, and computing technologies, is pushing us toward the deployment of fully autonomous AI-controlled weapons.

“What people are really worried about is that one day AI might control and even enslave humans,” Mr. Xia added.