``OpenAI will leave the EU if there is a full-scale regulation,'' CEO Sam Altman said


La Moncloa - Gobierno de España

OpenAI CEO Sam Altman said that if OpenAI is technically unable to comply with the provisions of the new AI law being prepared by the European Union, it may suspend service operations in the EU.

OpenAI Could Quit Europe Over New AI Rules, CEO Altman Warns | Time

Sam Altman Says OpenAI Will Leave the EU if There's Any Real AI Regulation


From May 23, 2023, Mr. Altman has been attending the OpenAI World Tour held in various countries such as Toronto, Rio, Lisbon, Warsaw, Paris, London, etc. He has a busy day, meeting with government officials from the United States.

As part of this, Mr. Altman took part in a panel discussion held at University College London. In the discussion, Mr. Altman revealed that he had discussed the AI law being enacted by the EU with the person in charge of the regulatory authority, and said that there was ``a lot of criticism'' about the current bill.

A particularly skeptical part of the bill is the definition of ``high-risk'' systems included in the bill, Altman said. Current language designates large-scale AI models like OpenAI's ChatGPT and GPT-4 as 'high-risk,' potentially forcing companies managing these models to comply with additional safety requirements. It seems that there is

Regarding these bills, Altman said, ``You either can solve these requirements, or you can't. has technical limitations,' he said. While showing a willingness to comply with regulations as much as possible, he has expressed the view that depending on the requirements, service suspension in the EU will be forced.

Mr. Altman has taken a position in favor of regulation of AI by law, and at a hearing held in the United States, he said, 'Government regulation to minimize the harm of AI while increasing its benefits.' Intervention will be important,” he

said . However, in the EU, where regulations such as personal information management are stricter than in the United States, if regulations that OpenAI's technology cannot handle are imposed, we will have no choice but to stop providing it.

Mr. Altman argues that such regulations on AI should be ``formed by diverse perspectives that reflect the public interest.'' As a concrete measure, we are looking for a team to conduct an experiment to establish a ``democratic process'' that can answer questions about the rules that AI should follow, and provide a grant of 100,000 dollars (about 14 million yen). We are launching a program through OpenAI to award 10 teams.

Democratic Inputs to AI

Regarding the 'democratic process', OpenAI describes it as 'a process in which representative people exchange opinions, hold deliberative discussions, make transparent decisions, and set out final results.' OpenAI is recruiting teams from around the world to develop proofs of concept for these processes.

``There are no intrinsic flaws, but subtle details matter,'' Altman said of the EU bill. At the same time, Mr. Altman said, ``I do not want regulations that limit users' access to technology,'' revealing the idea that he does not want to harm small businesses and the open source AI movement. Did.

in Posted by log1p_kr