Bachelor's degree programs in artificial intelligence and machine learning will have a dedicated set of ABET program criteria starting in fall 2027. ABET's Computing Area Delegation gave the new criteria final approval on November 6, 2026, according to CSAB, the ABET member society that leads accreditation work for computing disciplines. ABET's 2026-27 computing criteria contain program criteria for computer science, cybersecurity, data science, information systems and information technology, and none for AI or machine learning.

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

Who the new criteria cover

The criteria apply to computing programs that use artificial intelligence, machine learning or a similar term in their titles. Like every set of ABET program criteria, they sit on top of the general criteria rather than replacing them. A program still has to meet the general computing requirements, including the five general student outcomes and the general curriculum floor of at least 30 semester credit hours of computing topics, and then meet the AI and machine learning requirements in addition.

The added student outcome

The general computing criteria list five outcomes every baccalaureate graduate must be able to demonstrate, covering problem analysis, solution design, communication, professional and ethical judgment, and teamwork. The AI and machine learning criteria add a sixth: graduates must be able to apply AI theories, models and techniques to design, build and evaluate AI-based solutions to complex problems. Programs will need to show evidence that students attain this outcome in the same way they already document the first five.

What the curriculum must include

The core requirement is at least 42 semester credit hours, or the equivalent, of AI and computing coursework. Within that, the approved text names seven fundamental areas:

  • AI foundations, such as reasoning, heuristic search and knowledge representation
  • Programming, data structures and algorithms
  • Data and knowledge engineering
  • Machine learning, including deep learning
  • Design and implementation of AI solutions
  • AI system architecture, infrastructure, MLOps and workflows
  • Secure, ethical and responsible AI

Beyond those fundamentals, the coursework must include advanced AI topics that build depth on the foundation, and coverage of at least one application area in which AI is used. Separately, the criteria require at least 9 semester credit hours of mathematics and statistics, covering statistical inference and modeling, linear algebra, probability and optimization. As with all ABET program criteria, the requirements name topics rather than specific courses, so a program can distribute them across its catalog in its own way.

How it compares with neighboring computing criteria

The 42-hour requirement sits between the computing disciplines already listed in ABET's 2026-27 computing criteria. Computer science programs must offer at least 40 semester credit hours of computer science coursework and at least 15 hours of mathematics and statistics. Cybersecurity, data science and information technology programs each require at least 45 hours of discipline coursework, with cybersecurity and information technology requiring at least 6 hours of mathematics. The AI and machine learning mathematics floor of 9 hours is lower than computer science's 15 but names a specific set of topics, with linear algebra and optimization included alongside probability and statistics.

How the criteria reached final approval

The criteria followed ABET's standard two-step path. On October 24, 2025, the Computing Area Delegation released them as proposed criteria, and ABET printed them in the proposed changes section of its 2026-27 computing criteria, with public comments accepted until June 15, 2026. The Computing Accreditation Commission approved them for implementation at its July 2026 meeting, and the delegation's November 6 vote was the final step. The same meeting approved revised program criteria for information systems, also effective fall 2027.

What programs should do now

A program with AI or machine learning in its title that expects an ABET review in the 2027-28 cycle or later will be reviewed against these criteria. The practical first step is a curriculum map: list each of the seven fundamental areas, the advanced topics and the application area against the courses that cover them, and confirm that the hours meet the 42-hour and 9-hour floors. Programs considering a first review can find ABET's eligibility requirements on its website, and the full approved text is in CSAB's 2026-2027 approved criteria changes document. The approved wording will also appear in ABET's criteria document for the 2027-28 cycle once ABET publishes it.

Sources