Saturday, 12 September 2026

AI Is Changing Computer Science. Are We Changing the Classroom?

Dear Friends,

Six years ago, in 2020, I made a video when engineering colleges were introducing CSE with different variants such as DS, AI/ML, Core CSE, IT, etc. (https://youtu.be/puRhpSSzHJ4). In the same year, I made another video on “Three Types of Engineering Students—which type are you?” (https://youtu.be/z4P-9dDuQEU).

In 2020, I did not anticipate AI’s maturity to be what it is today. If you look at its brief history, in 2012, the AlexNet neural network crushed traditional computer vision benchmarks at the ImageNet Challenge. It proved that deep neural networks, combined with graphics processing units (GPUs), could extract patterns from raw data far better than human engineers could by hand.

In 2017, Google researchers published “Attention Is All You Need,” introducing the Transformer architecture. This sparked the birth of Large Language Models (LLMs). The public release of ChatGPT in late 2022 brought general-purpose AI reasoning into everyday consumer awareness almost overnight. If we want to summarize the journey: 2011: AI predicts → 2015: AI perceives → 2018: AI understands → 2020: AI generalizes → 2022: AI generates → 2023: AI collaborates → 2024: AI reasons and uses tools → 2025: AI acts → 2026: AI increasingly manages end-to-end workflows.

If I review my two earlier videos in 2026, the first video on the different varieties of CSE is no longer significant or relevant. However, the second video is highly significant.

Many of you know Srikanth Velamakanni, NASSCOM Chairperson, Co-Founder of Fractal, and an alumnus of IIT Delhi and IIMA. Fractal went public in February 2026, reaching a valuation of approximately USD 1.93 billion. I admire Srikanth’s LinkedIn posts. Interestingly, a couple of weeks ago, Srikanth posted about how the CSE branch is declining in the US and the UK.

The chart shows a clear reversal in the long-term growth of Computer Science in both the US and the UK. Using 2024 = 100 as the reference point, US CS enrolments increased from roughly 62 in 2015–16 to 100 in 2024, but then fell to about 91–92 by 2025–26, implying an approximately 8–9% decline from the peak. In the UK, CS applications rose from around 60 in 2014 to 100 in 2024, then dropped to about 86 by 2026, representing a decline of approximately 14%.

The downturn began soon after generative AI became mainstream, suggesting that AI-driven automation, weaker entry-level software hiring, technology-sector layoffs, and a shift toward AI, data science, and interdisciplinary programs may be influencing student choices. However, the chart shows correlation rather than proving that AI alone caused the decline.

You might be wondering: What is the status in India? There are no major statistics published yet, but as per my assessment, it looks something like this: 2024: 100, 2025: ~100–102, 2026: ~97–100. Hence, we seem to be witnessing a plateau or the beginning of a decline.

Srikanth argues that while the low-hanging fruit of basic software engineering is disappearing, this opens up tremendous new possibilities for solving much harder problems. And this raises a much bigger question: In an increasingly AI-driven world, what should colleges teach to help students build a 40-year career? He suggests that colleges teach critical thinking, learning to learn, problem-solving using first principles, judgment and ethics, empathy, compassion, entrepreneurship, and, most importantly, leadership.

India has approximately 1,300 universities and 65,000 Higher Education Institutions as of 2026. India has approximately 4.5 crore higher-education students; around 46 lakh (≈10%) are studying Engineering & Technology, while roughly 4.0 crore (≈90%) are studying non-engineering disciplines. Among engineering students, nearly 22 lakh (47%) are in CSE-plus-IT, AI/ML, and AI/Data Science-related branches.

If we don’t plan well, the demographic dividend could become a demographic curse. NITI Aayog estimates that another 8.7 crore youth aged 15–29 are outside education, employment, and training (NEET—Not in Education, Employment or Training), highlighting the enormous challenge of moving young people from outside education and employment into skills and productive livelihoods.

Time to think differently and teach differently in the classroom!

Ravi Saripalle

No comments:

Post a Comment