President Joe Biden issued the first-ever national security memorandum on artificial intelligence on Thursday, detailing goals for how the government should foster cutting-edge AI while also advancing international consensus around the powerful and rapidly evolving technology.

The White House warned that the US, which has been a global leader in artificial intelligence, can't take its advantage for granted.

"We are all familiar with past instances when we saw critical technologies and supply chains that were developed and commercialized here in the US migrate offshore for lack of critical public sector support," the document says. "That is why we are laser focused on maintaining the strongest AI ecosystem in the world here in the United States."

At the same time, the memorandum notes that US AI efforts must be governed by the "critical guardrails" established in 2023 by Biden's executive order on safe, secure and trustworthy artificial intelligence.  

The memorandum directs the National Economic Council to "coordinate an economic assessment of the relative competitive advantage of the US private sector AI ecosystem." It also notes that the country will need to maintain its advantage by investing in semiconductors, infrastructure and clean energy.

Since OpenAI released its ChatGPT generative AI chatbot in 2022, the promise and risks of artificial intelligence have been much debated. In its simpler forms, gen AI can help job-hunters refine their resumes or help students manage their time. It has more glamorous uses too. The technology, for instance, helped pull John Lennon's vocals out of an old tape for use in a new Beatles song. But many experts worry about the misuse of AI, which can create fictional images that look very real, such as deepfakes of singer Taylor Swift that falsely suggest she endorsed former President Donald Trump for re-election. 

AI also has the potential to shake up fields from science and medicine to software development, cars and military technology.

The full document contains about 38 unclassified pages, with a classified appendix, according to The New York Times. Some of its statements are obvious, The Times points out. For example, it states that AI systems must never be allowed to make decisions about using nuclear weapons.  

Biden is not running for re-election, and his term ends in January. It is unclear how or if the new president, whether Trump or Vice President Kamala Harris, will follow the policy spelled out in the document.  

"Today's NSM is just the latest step in a series of actions thanks to the leadership and diplomatic engagement of the president and vice president, and there will be additional steps taken in the coming months to further support US leadership in AI," the statement says.