Discussion

US justice department is exploring AI for university admissions investigations

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#14205

The US Department of Justice is exploring using AI to analyse university admissions data in investigations into alleged discrimination, former Civil Rights division head Harmeet K. Dhillon said at a Monday talk. The department has not specified which tools it is considering or how they would be used, leaving a consequential question about AI in civil-rights enforcement open.

Watch Desk analysis

What happened

The Crimson reports that Dhillon said the division had traditionally relied on statisticians to analyse admissions data, but staff were now exploring AI “within the bounds of government policy”. She said data arrives quarterly from dozens of universities and that she had opened more than 100 investigations. Those figures are Dhillon’s account, not an independently established count in the report.

The report says the remarks came during a Harvard Law School Federalist Society event. Dhillon had announced her resignation from the post less than two weeks earlier. She did not identify any AI system or describe its proposed role in the investigations.

Why it matters

Admissions reviews can affect institutions and students, so the methods used to identify patterns deserve scrutiny. AI could help staff sift large volumes of data, but the announcement does not establish that a system has been selected, deployed or used to reach findings. That distinction matters: exploring a tool is not the same as letting it decide a case.

The department’s existing investigations concern alleged discrimination, not findings established by this announcement. The Crimson also reports that Harvard has been among the institutions investigated and that a judge dismissed a DOJ lawsuit seeking admissions records on procedural grounds on 2 November.

Our read

This is an early signal of AI entering a sensitive government process, not evidence that an algorithm is already judging admissions. The useful next step is for the department to explain what it wants the technology to do, how staff would check its output, and what safeguards would apply. “Within policy” is a starting point, not a complete account of oversight.

What to watch

  • Whether the DOJ names the AI tools under consideration and the tasks they would perform.
  • How analysts would test for errors, bias and inappropriate conclusions.
  • Whether the department publishes safeguards or gives universities and students a way to challenge AI-assisted analysis.

Discussion spark: Before AI is used in a civil-rights investigation, should the DOJ have to disclose the system and its validation methods, or can those details remain internal?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.