Recommendations on law enforcement’s use of artificial intelligence are headed to the White House after the National AI Advisory Committee’s (NAIAC) unanimous vote on Sept. 4 in favor of the report containing findings and recommendations. 

Law enforcement agencies need to follow a standardized AI field testing checklist before integrating AI tools into normal use, according to the NAIAC Law Enforcement Subcommittee’s report released earlier this summer. 

“Very few resources are available to help guide the AI industry, law enforcement departments, and independent researchers through the process of testing AI tools when they are provisionally used in the field,” the report says, adding that it provides “infrastructure” for testing.  

The checklist will provide law enforcement officials with requirements they must meet while using AI in the field to test whether the tools can be “accurate, fair, high-performing, and cost-saving.” According to the report, this can prove difficult when being tested in “relatively sterile conditions” used to test new technologies typically. 

The first recommendation is for the Office of Management and Budget (OMB) to encourage Federal law enforcement agencies to follow a checklist when testing AI. A possible checklist drafted by the subcommittee may include evaluating the AI tool’s use of risk assessment, personal information and identification, and AI surveillance.  

A field-testing requirement can be waived if the agency’s “use policy restricts the tool’s use to the same use policy” and if under near same conditions “it has been previously field tested by another agency.” 

Additional recommendations include that OMB require that the field testing plans and testing results be made publicly accessible even if the AI tool isn’t adopted. The third recommendation states that funding and research support should be made available for state and local law enforcement to conduct AI field testing, with funding provided to the Bureau of Justice Assistance through special-purpose grants. 

The document follows President Biden’s executive order on AI that requires real-world testing of AI tools. The subcommittee published its initial draft of its field testing recommendation in May.  

 The recommendations will be submitted to the president and the National AI Initiative Office – established in 2021 as a hub for Federal coordination and collaboration in AI research and policy – for their adoption.  

Read More About
Recent
More Topics
About
Weslan Hansen
Weslan Hansen
Weslan Hansen is a MeriTalk Staff Reporter covering the intersection of government and technology.
Tags