This executive order aims to streamline multiple aspects of AI to protect workers and enhance the federal government's role in the AI landscape. Here's a breakdown of the key points from this significant development:
The executive order makes allowances for Washington to monitor private sector development of powerful AI systems.
Directives regarding AI will go out to more than a dozen agencies, they will have 90 to 240 days to comply.
A White House AI Council will be appointed to act as coordinator of government AI activities. The council will be staffed with reps from all major agencies, and the White House Deputy Chief of Staff for Policy will chair.
The order suggests serious concerns about the rapid advancement of AI tech to create sophisticated cyberweapons.
Companies working on powerful AI models must regularly provide reports to the Commerce Department detailing their plans to protect their technology from digital subversion.
Heavyweight cloud services providers like Amazon and Microsoft will be required to alert the government when foreigners rent server space to train these AI models.
Federal agencies will have 3 months to identify AI use risks in the industries they oversee and must create guidelines for how private companies in those industries will incorporate the AI Risk Management Framework within 6 months.
The executive order privacy protection guidelines around the data used to build artificial intelligence systems.
The order also says that the technology should not be used to “encourage undue worker surveillance” in the workplace.
The Department of Education will create an “AI toolkit” for the education sector to help them implement recommendations that have already been made for using artificial intelligence in the classroom.
The department will also create resources and guidelines that “address safe, responsible, and non-discriminatory uses of AI in education” within one year.
This is not an exhaustive list of what was included in the EO. There are guidelines included for labor, microchip manufacturing, immigration, and a host of other sectors. The general tone is that while AI can be incredibly useful there needs to be oversight and transparency in how and where it is used. You can read the full executive order here.