Federal agencies may not be on the same page when it comes to artificial intelligence programs.
A newly-released review outlines numerous artificial intelligence policies, programs and projects across the federal government but notes they may not be consistent with each other.
The report, issued by the Advanced Technology Academic Research Center (ATARC) AI & Data Policy Working Group, covers six agencies developing AI technologies on behalf of the entire federal government and another 10 agencies developing AI or policy specific to their departments.
“There are currently dozens of separate AI ethics, policy, and technical working groups scattered among various federal departments and agencies, spanning the defense, civil, and legislative spheres,” the report states. “Each of these groups is pursuing important goals of defining policies related to AI within their specific charters. While a few overall governance structures for federal AI policy have been put in place, we are concerned that the evolving policies may be incomplete, inconsistent, or incompatible with each other.”
The report does not make specific AI policy recommendations to agencies or the Biden administration. Rather, it provides “a general framework and an assessment of the current state of federal government AI policy,” with the stated intent of “ensuring that the resulting policies are aligned to the values and priorities of the American people, are mutually consistent, and collectively complete.
Numerous uses of AI and big data across several arenas are addressed in the document, including for mail delivery, the workforce, criminal justice reform, environmental efforts and military purposes. The report calls for “coordinated policy” across the federal government and participation among numerous federal agencies in creating a governance structure, which will promote accountability and ethical AI use.
“If federal agencies contribute through governance structures to approve and adopt a set of principles and tools as presented in this framework, they can hold one another accountable and share their data accordingly,” the report states.
The report highlights three issues its authors believe to be critical regardless of policy considerations
1. Constitutionality: Any set of AI policies should first and foremost be designed and implemented to support the values and goals expressed in our constitution, as amended.
2. Completeness: the federal government’s AI policies should address all areas of AI policy that are necessary to ensure that AI is developed for the benefit of the nation’s citizens.
3. Consistency: the federal government’s AI policies should define the nation’s AI policy goals, without confusion or conflict, when enforced as a whole, regardless of which agency or organization develops and implements them.