The White House is increasing pressure on American artificial intelligence (AI) providers to use them responsibly. Four top executives were summoned there on Thursday: Sam Altman, CEO at OpenAI, Dario Amodei, CEO at Anthropic, Satya Nadella, CEO at Microsoft, and Sundar Pichai, CEO at Google and Alphabet. They met US President Joe Biden, Vice President Kamala Harris, Secretary of Commerce Gina Raimondo and a phalanx of senior Biden associates. The tech execs had to listen to “concerns about the risks associated with AI.”
Biden underscored the fundamental responsibility of companies to ensure their products are safe before they are deployed or released. It is essential to reduce the current and potential risks of AI for individuals, society and national security. It is about security, human rights, civil rights, data protection, jobs and democratic values.
“Be exemplary!”
The invited managers should be role models and actively promote responsible innovation with appropriate measures to protect people’s rights and safety, the White House demands. The lack of transparency regarding AI towards authorities, politicians and the public, the importance of protecting AI systems against attacks and the importance of evaluating and verifying the security and effectiveness of AI systems was also discussed. In diplomatic English, such scolding is referred to as “frank and constructive discussion“. The four managers assured their future cooperation.
This includes participation in the public evaluation of AI systems as part of Defcon 31 in August. The companies Anthropic, Google, Hugging Face, Microsoft, Nvidia, OpenAI and Stability AI have confirmed. The evaluation platform programmed the latter. Thousands of researchers should be able to check to what extent the AI models of the participants adhere to the blueprint for a fundamental rights charter on artificial intelligence and a risk management framework for AI. Both are documents that the White House and the US standardization institute NIST have worked out as part of.
Money for 7 new research facilities
In addition to criticism of the industry, the US government also has carrots on offer: shortly before the meeting, the National Science Foundation announced $ 140 million for seven new AI research institutes at. With research and development, they should contribute to breakthroughs in climate research, agriculture, energy, health, education and IT security. In total, there will be 25 AI research institutes in the United States dedicated to advancing AI that is “ethical, trustworthy, responsible, and committed to the common good.”
Biden also wants to sweep his own house. A draft directive will be publicly consulted in the summer; its purpose is to help US agencies develop, procure, and deploy AI systems while protecting people’s rights and security. The White House hopes that companies, US states, municipalities and others will follow these guidelines as a result.
(ds)
