Programs preparing an ABET self-study report can use artificial intelligence tools to help assemble it, but ABET's published statement on AI draws firm lines around the evidence. AI may help gather, organize and analyze. It may not change the data, and it may not stand in for the faculty judgment that the continuous improvement criterion depends on. With self-study reports for programs reviewed in the 2027-28 cycle due next summer, many program coordinators are deciding now how far to let these tools into the drafting process.

Credential Review is an independent publication. It reports on ABET's policies and does not speak for ABET.

What the statement covers

ABET's statement on accreditation and artificial intelligence technologies addresses AI in quality assurance work: how institutions, volunteer evaluators and ABET's own staff use AI tools in the accreditation process. It does not regulate what programs teach about AI or how students use it. ABET says it does not prescribe whether or how institutions govern AI or use it in student learning, and notes that AI-related topics may be added to its criteria in the future. The page as published carries no effective date or revision date.

What programs may use AI for

The statement lists three permitted uses for institutions and programs:

  • Gathering and synthesizing information for an accreditation review
  • Supporting the analysis of assessment data used for continuous improvement
  • Helping develop and organize accreditation materials

In practical terms, that covers tasks such as collating course syllabi, organizing assessment results into tables, summarizing advisory board minutes for the program's own use, and drafting the structure of a narrative section.

Where the lines are

Two prohibitions apply to programs. First, AI must not alter, manipulate or misrepresent collected data or documented evidence. Second, AI cannot replace human judgment in continuous improvement. The second matters because ABET's continuous improvement criterion asks programs to evaluate assessment results and act on them; under the statement, the evaluation and the decisions that follow have to be made by people.

The statement also sets out responsibility. Human oversight is required, qualified personnel must verify that submitted information is accurate, and only authentic, verifiable data may be presented as factual. Programs should acknowledge their use of AI, and the institution must attest to the accuracy of the narrative and the data it submits.

What it means for evaluators

For the volunteer program evaluators who review self-studies and visit campuses, the statement says AI shall not replace human judgment in any part of the evaluation process, and that review materials and deliberations remain subject to ABET's confidentiality policies. The statement does not separately address whether evaluators may load review materials into AI tools; the confidentiality policies are the controlling rule. For ABET staff, AI used in administrative processes must be built on verified data and evaluated for accuracy and bias before and after it is put in place.

Why the attestation matters

A self-study is the record a review team works from, and ABET's procedures give it consequences. After a visit, ABET's accreditation policy and procedure manual gives an institution seven days to report errors of fact in the team's preliminary findings and 30 days after the draft statement to respond on shortcomings, and a shortcoming is only treated as resolved if the correction is documented with signed official records. A narrative or data table that a tool produced and nobody checked is exactly the kind of error those steps are slow and costly to repair. The attestation places responsibility for accuracy on the institution regardless of how the document was drafted.

A working approach for a self-study team

  • Decide in advance which tasks AI tools will assist with and record that decision, so the acknowledgment ABET asks for is easy to write.
  • Keep the source data, such as rubric scores, survey results and course records, in its original form and generate every table from it, so any figure in the report can be traced back.
  • Have the faculty member responsible for each outcome review and sign off on that section, including any interpretation of results.
  • Check what institutional policy says about putting student records into outside tools, since privacy rules apply independently of ABET's statement.
  • Read the current statement on ABET's site before submission, since it may be revised.

Sources