For the first time, US President Biden wants to introduce detailed guidelines for the use of artificial intelligence in government agencies. As David E. Sanger reports in the New York Times, a national security memorandum will set “guardrails” for the use of AI tools – for example, in decisions about nuclear weapons or asylum applications. The new guidelines are intended to help speed up government processes using AI while limiting potential risks. National Security Advisor Jake Sullivan highlights the particular challenge that the U.S. government owns or funds only a few of the key AI technologies. Unlike earlier technological developments such as nuclear physics or the Internet, the private sector is now the driving force behind AI innovation.