Artificial intelligence sits inside computer science. The Computer Science Curricula 2023 report (CS2023), written jointly by the Association for Computing Machinery (ACM), the IEEE Computer Society, and the Association for the Advancement of Artificial Intelligence (AAAI), lists AI as one of 17 knowledge areas of the discipline. It gives AI 12 of the 162 core hours that every computer science graduate is expected to cover. As of October 2026, CS2023 is still the latest edition of these guidelines.
Key takeaways
- Computer science is the broader field, and AI is one of 17 knowledge areas in the 2023 joint curriculum from ACM, IEEE Computer Society, and AAAI.
- AI covers 12 knowledge units, including search, machine learning, planning, natural language processing, robotics, and computer vision.
- AI depends on other areas: the curriculum counts hours shared with algorithms and with mathematics and statistics.
- The report notes that AI, data science, security, and software engineering have been separating from computer science as distinct disciplines.
- CS2023 remains the current edition in October 2026; the ACM curriculum site lists no successor.
How the 2023 curriculum defines AI inside computer science
The Computer Science Curricula 2023 report (DOI 10.1145/3664191) describes AI as the study of problems that are difficult or impractical to solve with traditional algorithmic approaches. Those problems often resemble tasks thought to need human intelligence, and the solution strategies tend to generalize over classes of problems rather than one case. The report states that AI techniques now run through everyday computing, including email, social media, photography, and virtual assistants. Computer science is the larger body of knowledge. The report organizes it into 17 knowledge areas, from algorithms and operating systems to security and software engineering.
Is CS2023 still the current edition in 2026?
Yes, as far as public sources show. The ACM curriculum site records endorsement by ACM on January 18, 2024, by the IEEE Computer Society on January 22, 2024, and by AAAI on February 22, 2024, and its pages checked on October 27, 2026 announce no revision or successor. The previous edition dates from 2013, which makes CS2023 the first full revision in a decade. No source consulted states when the next one is planned, so any date for it would be a guess. Readers comparing the table below with a 2026 course catalog should treat the hour counts as recommendations from 2024, not as requirements that any school must follow.
Where the hours go: selected knowledge areas
The report assigns each area a count of CS Core hours (topics for every computer science major) and KA Core hours (extra topics for students studying that area in depth). The figures below are copied from its body-of-knowledge table. Across all areas the totals are 162 CS Core hours and 270 KA Core hours.
| Knowledge area | Knowledge units | CS Core hours | KA Core hours |
|---|---|---|---|
| Artificial Intelligence | 12 | 12 | 18 |
| Algorithmic Foundations | 5 | 32 | 32 |
| Mathematical and Statistical Foundations | 5 | 55 | 145 |
| Software Development Fundamentals | 5 | 43 | not listed |
| Data Management | 13 | 10 | 26 |
| Security | 7 | 6 | 35 |
Twelve of 162 is about seven percent of the common core, so a graduate who meets the minimum has spent a small slice of the degree on AI. Mathematics and statistics carry more than four times as many core hours (55 against 12), which fits the way AI methods rest on probability, linear algebra, and optimization.
What AI borrows from the rest of computer science
The AI chapter lists 12 knowledge units: fundamental issues, search, knowledge representation and reasoning, machine learning, applications and societal impact, logical reasoning, probabilistic reasoning, planning, agents and cognitive systems, natural language processing, robotics, and perception and computer vision. Several units lean on neighboring areas: three core hours on uninformed search are counted under Algorithmic Foundations, and one hour on probability is counted under Mathematical and Statistical Foundations. The report also says the AI area connects to data science through the Data Management area and a sample data science course.
A rule-based problem and a learned one (illustrative example)
This is an illustrative example, not a case from the report. Sorting a list of numbers is fully specified: any correct algorithm gives the same answer, and a student can prove it works. Deciding whether an email is spam has no such specification, because senders change tactics and people disagree on borderline messages. A common approach is to train a model on labeled examples and measure its error rate on new mail. Both tasks belong to computer science. The second falls in the AI area because the solution is learned from data and judged by measured performance instead of a proof.
Why the labels have been separating
The report observes that computer science has been rapidly fragmenting, with software engineering, data science, security, and more recently AI emerging as separate disciplines. The report calls this an opportunity, since institutions can share costs, and a challenge, since the disciplines need clear differences for students to choose between. Options range from certificates and minors to separate majors. The report also changed the area's name from Intelligent Systems in the 2013 guidelines to Artificial Intelligence, citing the most common terminology, and it raised the emphasis on neural networks and representation learning while keeping search and symbolic methods.
FAQ
Is AI a different field from computer science?
Under the 2023 joint curriculum, AI is a knowledge area within computer science. Some universities offer stand-alone AI degrees, and the report acknowledges this separation, but the curriculum itself places AI inside the discipline.
Can someone work in AI without a computer science degree?
Possibly, since statistics, mathematics, engineering, and other fields feed into AI work. The curriculum report describes degree content, not hiring rules, so it does not settle who is qualified for a given job.
Does a computer science degree teach machine learning?
The curriculum allots four CS Core hours and six KA Core hours to machine learning within the AI area. Whether a given program covers more depends on the courses that institution offers.
Has a newer curriculum replaced CS2023?
Not according to the ACM curriculum site as of October 27, 2026. CS2023 is the current joint guideline, and no successor or revision has been announced there.
Which one should a student choose?
The report leaves that to institutions and students, noting that they can focus on different competency areas such as software, systems, or applications development. Course catalogs and prerequisites at the specific school are the practical guide.
Related reading
- Artificial versus automated intelligence
- The role of computer vision in AI
- How conversational interfaces developed
- AI in research and discovery
More in AI Basics.
Last verified: October 2026
Sources
- Computer Science Curricula 2023 (CS2023), ACM, IEEE Computer Society, and AAAI joint task force, 2023; DOI 10.1145/3664191
- CS2023 endorsement dates: ACM January 18, 2024; IEEE Computer Society January 22, 2024; AAAI February 22, 2024