The survey looks at the influence of AI on the study habits, class choices, and career prospects of Harvard students, with responses from 326 undergraduates
Use
A striking 87.5% of students have embraced generative AI, with an overwhelming 95% using ChatGPT, while Claude and GitHub Copilot (a programming assistant) are each used by around 20%. A significant 30% of users are also investing in premium AI subscriptions, demonstrating a willingness to pay for the edge AI provides. 25% of students feel that AI has lightened their academic load, reducing their need ask course staff for help, even completing readings, though it's interestingly not keeping them away from lectures. 20% of students have already changed their course paths due to AI's influence, and more than half want Harvard to step up and offer more courses that delve into the profound future impacts of AI.
What are they using AI for?
For one third AI is replacing Wikipedia and Google search
Help with writing assignments (coming up with idea, drafting, proof-reading)
Writing emails
Helping with programming assignments
Data processing
Beliefs and fears
None of the above surprises me but it is the detail where the most interesting stuff is to be found. Students are increasingly anxious about academic fairness, fearing that others may gain an unfair edge through AI, while those from less privileged backgrounds feel uneasy around the high costs of premium AI tools. Although those without financial aid are twice as likely to pay for costly AI subscriptions, widening the gap between them and their peers who receive partial or full financial assistance.
45% of the students are concerned that AI could hurt their future career prospects. But the most fascinating belief to emerge from the data is that many are wrestling with a sense of purposelessness in their education as AI advancements accelerate at a breakneck pace. A staggering 40% believe AI might outstrip human abilities within the next 30 years, a belief that underscores their deep-seated fears about the long-term consequences of AI.
What do they want?
Facilitate access to AI with free access to a paid plan of ChatGPT or Claude.5 Establish & enforcing consistent rules on AI use
Provide AI-aware career planning services
Offer courses exploring the future impacts of AI
Help students find meaning in education and beyond
Conclusion
The impact on undergraduates is clear and cannot be ignored. They’re using it. That use will not go away. It is often said that engagement is a problem in learning. That appears to be true in this case but the lack of engagement is with faculty and administrators.
If we listen to learners we can guide its use towards good outcomes. The sly use of AI has clearly become normalised. This will continue for some time but it is now time for institutions to step up and get it integrated into their teaching. I can’t help but ruminate over the fact that students are already thinking about the futility of learning skills that will be done better by AI. It’s their future and they are concerned that their very expensive education may not be worth the effort.
Meanwhile we need to get going, This means creating the future, not just letting it happen to us. Think about what institutions need to do in this future. These students are giving us clear recommendations. Above all, take AI seriously. Use it to teach and learn. Above all, get it into the curriculum so we all know what we are dealing with.
PS
On costs the lifeboat seems to be righting itself. For 70% AI used is free, another 4.6% are paying 0-$15, 24.3% paying what looks like the standard $20. This is modest for students at Harvard and services are some good services are now free. Compared to the cost of textbooks ($1000-$1200) a year) and other services at Harvard, it seems reasonable. Interestingly some Universities, like the University of Michigan, are paying for all student licences. It will happen over time, early days.
No comments:
Post a Comment