The U.S. intelligence community (IC) has a new rulebook to guide its use of open-source data and publicly available information, including artificial intelligence (AI)-assisted services and generative AI (GenAI) tools.
On Dec. 2, the Office of the Director of National Intelligence (ODNI) unveiled the “Intelligence Community Standard” (ICS) framework which outlines how intelligence agencies can use “publicly available information, commercially available information and open source intelligence.”
According to the ODNI, the ICS comes in response to the explosion of open-source data and reporting over the past decade, which has drastically expanded the intelligence landscape.
“The rapid and continued growth of Publicly Available Information and Commercially Available Information, alongside the increasing utility of [AI] necessitate an updated and forward-looking approach to the use [these resources] in IC products and reports,” the document reads.
The new standards lays out how intelligence agencies should describe those sources in their reporting and analysis. It establishes standards for reference and citation in “disseminated analytic products” and “covered analytic products” to include all source products, multi-intelligence products, standalone open source intelligence analytic products, and other standalone open source intelligence products.
The ICS also provides detailed guidance on describing sources, including assessing their credibility, and emphasizes the need for agencies to minimize the sharing of “unevaluated” open-source data.
Additionally, the standards outline how intelligence officers should cite AI-generated reports and services, including classification models and GenAI systems. According to the ICS, these guidelines are crucial to how the IC can leverage and harness AI and machine learning, particularly for open source intelligence.
In addition to citation and descriptions, the ICS outlines how the retention periods for externally sourced material cited in intelligence products may differ from those of internally created products, which are often kept permanently.
The document also notes that emerging technologies may not always fit into the categories outlined in the standards. The framework also lays out plans for training the IC workforce on guidelines outlined in the framework.
