US President Joe Biden is expected to sign an executive order on the regulation of artificial intelligence (AI), the Washington Post said on October 24.
The order, likely to be signed on Monday, October 30, would have at least three objectives. First, the order will require advanced AI models to be assessed before they can be used by the U.S. government and its federal employees. It is unclear to which models this rule might apply and what criteria the assessment might include.
Second, the order will direct select U.S. agencies, including the Department of Defense, the Department of Energy and certain intelligence groups, to study how they can use AI to improve national cybersecurity.
According to the Washington Post, the executive order will also make it easier for skilled technology workers to immigrate to the US.
The executive order may include a fourth part
Separate reports from Semafor dated September 27 suggest the order will require cloud companies to report how tech companies use cloud computing resources. This rule is intended to detect foreign AI projects that could pose a national threat.
Although the Wall Street Journal’Administration sources did not comment on that aspect of the executive order, reports from Politics on October 12 suggest this is still on track.
Some critics have raised concerns that any surveillance effort focused on quantity-based monitoring of cloud computing resources could impact other non-AI industries, including game development and Bitcoin and cryptocurrency mining.
The details of the order, while supported by many sources, are not yet finalized. Current reports indicate that details and timing are subject to change.
In addition to issuing the order itself, Biden and the White House will reportedly host an event on October 30 titled “Safe, Secure and Trustworthy Artificial Intelligence.” The US president previously addressed the UN on AI regulation in September.