People affected by consequential government decisions were given a place in the White House’s October 12, 2016 artificial intelligence agenda. The administration released a report recommending additional care when federal agencies use AI to make decisions about individuals, or to help officials make them. The recommendation concerned both effectiveness and fairness, not simply whether a system could produce an answer. [ai-release][ai-report]

The report’s Recommendation 16 called for verification and validation based on evidence. Its scope included decision support: keeping an official involved would not, by itself, place a system outside the proposed safeguard. For someone affected by a decision, that distinction matters because the recommendation follows the use of AI, rather than only fully automated outcomes. [ai-report]

Recommendation 17 addressed a different route into public services: federal grants to state and local governments. It urged agencies to review grant terms so that AI products or services bought with those funds would deliver sufficiently transparent results and have evidence of effectiveness and fairness. That was a proposed condition on public spending, not a finding that funded products already met it. [ai-report]

Our public-interest reading is that these recommendations assign work to the institutions choosing and financing systems, rather than asking affected people to accept a technical promise. They do not establish that testing was completed or grant terms changed. The announcement and report come from the same government policy process, not independent assessments of success; implementation evidence would be needed to judge whether the proposed protection reached anyone. [ai-release][ai-report]

Source documents

  1. The Administration’s Report on the Future of Artificial Intelligencehttps://obamawhitehouse.archives.gov/blog/2016/10/12/administrations-report-future-artificial-intelligence
  2. Preparing for the Future of Artificial Intelligencehttps://obamawhitehouse.archives.gov/sites/default/files/whitehouse_files/microsites/ostp/NSTC/preparing_for_the_future_of_ai.pdf