Mistral, Europe’s leading artificial intelligence developer, and other market players accuse U.S. rivals of abusing security concerns to cement their dominance. Europeans reject Anthropic’s call for advanced model developers to slow down development, writes Reuters.
Photo credit: Steve A Johnson / unsplash.com
Human Dario Amodei General Manager called Colleagues have slowed improvements to artificial intelligence models and advocated relaxing antitrust regulations to give advanced labs a chance to work together to develop safety measures. The need for this became public when researchers warned that artificial intelligence could wipe out the human race. The initiative has the support of SpaceX CEO Musk and OpenAI CEO Sam Altman; meanwhile, European companies and officials have taken a more skeptical stance. “Some existing market leaders are using this moment to consolidate their position, calling for regulation to gain an advantage over their competitors.”France’s Mistral claimed, adding that the threats had been known for months.
There are not many developers of advanced AI models in Europe; the region’s main hope for creating models on the level of the United States and China is Mistral, which was created three years ago. In fact, Europe actively uses OpenAI, Anthropic or Microsoft models, while still relying on American technology. Proton said Anthropic’s move was driven by the company’s self-interest and the desire to maintain users’ dependence on its services; while they support Amodai’s idea of incorporating an artificial intelligence lab into the mix, France-US Hugface said development doesn’t need to be slowed down, but accelerated. “Independent Appraiser”.
Leading artificial intelligence developers have long warned that these technologies pose an existential threat to humanity and that protective measures are needed to stop systems that could cause harm. Dario Amodai emphasized that the development of artificial intelligence is ahead of the pace of government regulation; leading developers should voluntarily agree to safety standards and allow permanent independent expert supervision to ensure compliance. In Europe, the proposed measures are considered excessive: a plan in which a few large companies themselves determine what is dangerous, choose inspectors and require exceptions to antitrust laws can hardly be called mature regulation.
The European Commission believes that existing AI-related regulations in the region enable Europe to play a key role in addressing risk minimization issues. The representative of Germany added that Europe urgently needs to achieve digital sovereignty and overcome technology gaps.
If you find an error, select it with your mouse and press CTRL+ENTER.