The 21st Century Red Telephone: When Superintelligence Rebels, We'll Call Them

Expectations for artificial intelligence are high, as are fears. Critics and industry leaders, not just from Silicon Valley, warn of existential dangers, or even the extinction of humanity. Optimists counter that technological progress cannot be stopped by any law in the world. The question therefore arises whether it is at least possible to create effective protective barriers. The leading powers, the US and China, seem to want to take steps in this direction, without anyone being able to guarantee it.

During Chinese President Xi Jinping's visit to U.S. President Donald Trump in Washington, the two leaders focused on the risks and opportunities of the new technology on Friday. The White House said the two sides agreed to establish a bilateral communication channel on artificial intelligence issues. They also agreed to create a peace-building forum for dialogue on "superintelligence," which is set to meet for the first time by November.

See more articles from iGuRu.gr when you search for news on Google.

iGuRu.gr preferred source

The choice of words is notable: at the request of Donald Trump , the leaders agreed, according to the administration’s headquarters in Washington, to speak of Superintelligence (SI) instead of Artificial Intelligence in the future. Trump, who is sometimes seen as a regulation skeptic, wanted to see the euphemism introduced at the UN General Assembly, as he sees the new name as a more appropriate description of the technology’s capabilities.

Restricting Artificial Intelligence is unlikely, of course, but the agreement at least shows a shared need for control.

The agreed communication channel is reminiscent of the Cold War-era Red Hotline, which was intended as a direct line of communication between Washington and Moscow to prevent accidental nuclear escalation. Recent reports that an AI malfunction in the US nearly triggered a direct military conflict with China highlight why this step is necessary. In this context, a crisis hotline seems essential for intervening in an emergency before autonomous systems can make decisions.

The urgent need for trusted agreements is demonstrated by recent incidents such as the Hugging Face attack , in which AI systems autonomously exited their test environments. OpenAI recently informed several companies, authorities, and universities that their websites may have been affected by out-of-control AI models. Such events make it clear that in addition to technological know-how, economic and political power is required to avoid losing control altogether.

newsbot

Although the press releases will range from very select to rare, I said I'd pass...because sometimes the editors hide.