Between 2024 and 2025, enrollment in CS as an undergraduate major at U.S. four-year universities declined 11%, mirroring a labor market where employment among the youngest software developers has fallen since 2024. But because degree completion lags enrollment by several years, the full effect will take time to surface. Meanwhile, AI-relevant majors keep gaining: AI-software master's graduates rose 17% from 2023 to 2024 (and 82% since 2022).

PhDs: all the growth went to academia

• New AI PhDs in the U.S. and Canada rose 22% from 2022 to 2024, but the share entering industry stayed flat — all of the growth went to academia, challenging the 'brain drain' narrative.
• Industry still takes the majority: about 65% of AI PhD graduates went into industry, down from a 77% peak in 2022, while the academic share has nearly doubled since 2022.
• Most AI-related graduate students are non-residents: 67% of AI-software master's graduates are nonresidents — a share now at risk from tighter U.S. student-visa policy.

Four in five students already use AI

• In Chegg's 2025 survey of university students across 15 countries, 80% said they had used generative AI for their studies — double the 40% reported in 2023. Usage ranges from 95% in Indonesia to 67% in the U.S. and U.K.
• Students lean on AI for higher-order work: understanding a concept (56%) is their top use, followed by researching (52%) and generating first drafts (46%). Anthropic found students use Claude mostly for creating (39.8%) and analyzing (30.2%).
• Of students who use AI for school, 56% input a question at least once a day; among non-users, the top deterrents are fear of misconduct accusations (45%), accuracy concerns (38%), and restrictive school policies (33%).

The share of U.S. public high schools offering foundational CS rose from 35% in 2017–18 to 60% in 2023–24, holding at 60% in 2024–25. Access is deeply uneven by school size — 91% of large high schools offer CS versus just 44% of small ones — and only 6.1% of students were enrolled in CS in 2024–25.

The implementation gap

• An estimated 50%–84% of middle- and high-school students now use AI for school tasks, yet only about half of middle and high schools have an AI policy — and just 6% of teachers say those policies are clear.
• Standards are thin on AI: of 45 states with K–12 CS standards, 29 include AI only minimally and 10 not at all; just six states have standards with significant AI-specific content (plus two in draft).
• AP Computer Science — a primary pathway to advanced CS — still contains no AI-specific content, and there are no state-level standards or funding for AI teacher training. Policy, curriculum, and training all have gaps.

Going global

• In 2025, about 93% of the world's countries taught CS, and more than 90% offer it to primary or secondary students — yet only 30% mandate it.
• Two countries moved decisively on AI itself: China (Beijing, Guangdong, Hangzhou) and the United Arab Emirates both mandated AI education starting with the 2025–26 school year, with grade-progressed curricula from literacy to system design.
• Others integrated AI into teaching rather than teaching about it — South Korea launched then reversed primary-school AI textbooks, while Estonia's AI Leap 2025 is piloting AI tools with 20,000 students and 3,000 teachers.

LinkedIn's relative AI skill-penetration rate measures how prominently AI skills appear in a country's member profiles versus the global average. India leads at 3.0 — almost three times the global average — followed by the U.S. at 2.0 and Germany at 1.8. A persistent gender gap shows up everywhere: in India, men list AI skills at 3.1 versus 1.9 for women.

Literacy is outpacing engineering — with exceptions

LinkedIn's new AI Skills Diffusion Index separates AI literacy (familiarity with AI tools) from AI engineering (building and deploying systems). In most countries, including India and the U.S., literacy skills are growing faster than engineering skills. The exceptions are the United Arab Emirates, Chile, and South Africa, where engineering-oriented AI skills show steeper growth since 2022. In the U.S., the fastest-growing literacy skills were AI prompting and Microsoft Copilot Studio; the fastest-growing engineering skills were AI agents, AI productivity, and AI strategy.