Donald Trump refuses to slow down the race for artificial intelligence. “The only control or ‘guide’ AI needs is a strong, intelligent (high IQ!) president,” he wrote on Truth Social. The American president assures that the United States is ahead of China in this area and describes calls to slow down its development as an “unhealthy plot”.
However, this release comes after a series of warnings from the AI players themselves. Dario Amodei, boss of Anthropic, called for slowing down its development in order to better measure the risks linked to the new capabilities of the models. He fears that, within six to 12 months, AI agents could take control of large portions of the Internet through networks of hacked computers. His concerns were relayed by Sam Altman, boss of OpenAI, and Elon Musk.
Evan Hubinger, researcher at Anthropic, goes even further: he estimates the risk that AI could cause the disappearance of humanity in the next ten years is more than 10%. However, this is a personal estimate and not a scientifically established probability. According to the “New York Times”, 1,386 employees of companies specializing in AI have also signed a petition to warn of the dangers of the technology. Feared scenarios include the creation of deadly infectious agents or access to nuclear weapons.
Governments called to intervene
Concerns are also affecting the industry. Monday in London, Patrice Caine, CEO of the French group Thales, estimated that companies could not take sole responsibility for containing AI. The technical means exist, but must be accompanied by rules imposed by the public authorities. “We need managers, and this is clearly the responsibility of governments.”
Same warning from the UN side. The High Commissioner for Human Rights, Volker Türk, calls for “urgent action” in the face of “unprecedented risks”. Faulty or hijacked autonomous systems could, according to him, paralyze essential services, disrupt communications or destabilize critical infrastructure.
He also worries that the most powerful models are controlled by a handful of mainly American and Chinese companies. “No country can regulate this technology alone. No company should be able to decide alone what risks the world should accept,” he warns.
Wall Street is scared
This accumulation of warnings even shook the markets. On Monday, Intel fell by 7.4%, Micron Technology by 6.4% and Nvidia by 3.9% on Wall Street, while cybersecurity specialists progressed.
Behind the debate on the dangers of AI is mainly the technological battle between Washington and Beijing. Slowing down would make it possible to better control risks, but could also offer a strategic advantage to China. Beijing also shares certain concerns: the head of Chinese intelligence has warned that AI could be used by foreign powers to destabilize the country.
