The Department of Homeland Security has issued a directive establishing DHS policy for the use and acquisition of artificial intelligence as part of efforts to advance AI innovation and governance while managing risks.
The directive issued on Jan. 15 states that specifications for technical requirements as required by AI acquisition contracts should address testing and evaluation requirements and risk management considerations.
According to the directive, such specifications should ensure that technical requirements support transparency, performance evaluation and improvement in AI; assess sustainability and environmental efficiency; and address data ownership and management.
What Are the Prohibited Uses of AI at DHS?
The new DHS policy outlines several AI uses that are not allowed within the department, such as relying on outputs of AI as the sole basis for law and civil enforcement actions or denial of government benefits and using data associated with the use of AI to make or back decisions based on improper consideration of race, gender, ethnicity, religion and medical condition, among other factors.
The directive also prohibits the use of AI in unlawful or improper systemic, indiscriminate or large-scale monitoring, surveillance or tracking of individuals.
DHS said improper targeting, profiling or discriminating against any person based on race, general identity, national origin and other individual characteristics is also prohibited.
Responsibilities of DHS Officials
According to DHS, the department’s chief AI officer is responsible for leading governance and oversight processes and structures and establishing strategic priorities for AI deployment efforts.
The directive states that the chief information officer supports the use of AI at DHS through a data management lifecycle framework and data standards and requirements. The CIO should also ensure that AI use complies with cybersecurity requirements and aligns with the department’s customer experience commitment.
Register now to join the Potomac Officers Club’s 2025 Artificial Intelligence Summit on March 20, to hear more about cutting-edge AI innovations from industry and government experts.
