Policy Change Lets DHS CAIO Waive AI Limits – MeriTalk

The Department of Homeland Security revised its artificial intelligence policy to allow its chief AI officer to waive requirements of the department's AI directive, MeriTalk reported. The revised document is a departmental directive, referred to in the quoted waiver language as "this Directive." Under the change, the chief AI officer may waive "any requirement of this Directive and implementing policy, including prohibited uses of AI at DHS." The authority is conditioned on exceptions remaining consistent with law, regulation and governmentwide policy.

The same update encourages the sharing and reuse of AI assets and prioritises technology developed or produced in the United States. MeriTalk does not name the department's chief AI officer, state when the revised policy takes effect, identify which uses appear on the department's prohibited list, or say whether a waiver has been issued under the new authority. The directive's full text is not reproduced in the report.

The report presents the change as a revision of DHS's existing AI policy rather than a new statute or regulation, with the exception expressly bounded by law, regulation and governmentwide policy. The quoted language extends the waiver to "implementing policy" as well as to the directive itself. MeriTalk's account is limited to three elements: the waiver provision, the encouragement of AI asset sharing and reuse, and the preference for technology developed or produced in the United States.