How universities and students are tackling the implications of AI use is having far reaching implications for the delivery of education. Higher education institutions seem to swing between two extremes; either they allow AI, or its use is considered cheating.
Education’s confusion with AI
This creates conflicting messages. Some are completely rethinking how they assess students, with an increase in in-person exams, reversing a trend kick-started by COVID, or redesigning coursework. They know that in many ways they’re fighting a losing battle; in one study, 94% of AI submissions into an examination system went undetected (and secured grades on average half a grade boundary higher than non-AI submissions). In other instances, countries are changing higher education itself, such as China’s overhaul of university courses around AI.
This is all playing out against a backdrop of questions around the role of higher education generally, amid concerns regarding value for money and how well students are being prepared for future careers. In the US, there are concerns that a decline in teenagers graduating high school starting in 2026 and projected to continue to 2041 will create an ‘enrollment cliff’ for colleges, with implications for their long-term viability.
Students and the rise of AI guilt
Where does this leave students? Confused and caught in the middle. There are reports of a growing sense of ‘AI guilt’, where students use the technology, don’t get caught, but feel like they are cheating. Jeff Sharlet, a professor at Dartmouth, noted in a social media thread that many students reported hating AI, but feel like they have to, with none describing it as beneficial for their education (interestingly, that’s reflected in the workplace too; 42% of AI-using employees say using AI feels like cheating).
At the same time, employers are expecting students to graduate with AI-relevant skillsets. One study found that 75% of decision-makers expect a degree of AI proficiency to become the standard across many non-technical roles in the next two years. A World Economic Forum report revealed that “Demand is increasingly moving toward applied and human-centric skills such as critical thinking, problem-solving and the ability to work effectively with AI agents.”
It’s something we can attest to; when we hire, we increasingly look at how people think rather than what they can produce. Anyone can generate a polished output now, so what separates candidates is the quality of the reasoning behind it. Someone who has outsourced their thinking throughout their education tends to show it quickly in an interview, and even faster on the job.
What this all means for students is that it’s confusing: on the one hand we’ve got universities giving out mixed messages, and on the other students not knowing which way to go. There is a thread running through it all: how AI is used (or perceived to be used). Universities that see it as cheating expect it to be used as a replacement for students’ thinking, and it’s a similar theme in Sharlet’s thread: there’s a sense that by using AI, students have given something up.
There is more than one way to use AI
What all this presupposes is that there is only one way to use AI: to do the work for the user. That’s certainly how many AI tools are positioned: taking away the drudge work so people can focus on the creative, or value-adding, tasks humans excel at. But in the context of higher education, the mindset seems to be that using AI equals outsourcing the learning. Completing reading assignments, writing essays and lab findings; these are some of the fundamental building blocks of many degrees. If that’s the only way to use AI, then it’s clear why people are racked with ‘AI guilt’.
The reality is that there are ways to use AI, in higher education, that don’t make people feel guilty or leave institutions accusing them of cheating or plagiarism. Ways to not diminish learning or undermine their ability to develop their skills. Ways in which AI can enhance the learning experience, and perhaps even relieve some of the pressure off universities and professors.
AI that genuinely augments the learning experience
For instance, could an AI chat bot be used to help students prepare for tests? Trained with course data, understanding where the student is on their learning journey, and delivering tailored quizzes, suggesting areas to work on, even producing mock papers that push and support the student in ways that work best for them. Right now, students are turning to generic models that might provide overviews, but they won’t be specifically relevant to that learner’s course or module, so the experience will be superficial at best.
These assistants wouldn’t be providing answers. Instead, they’d be prompting students to go deeper with their learning, challenging them in the areas that they need focusing on. Restricted finances and the realities of today’s education market mean most universities can’t provide one-on-one tutoring, at least not with professors; the AI assistant would be a step back towards that tradition.
This is AI for personalized learning. It’s not doing the work for students. It’s not thinking for them. They’re getting individual study materials that are tailored to their strengths and weaknesses and, importantly, on their exact courses. Not a generic idea of what a module in Old English or property law or physiology of cardiovascular and respiratory systems is, but the actual courses themselves.
Reframe how education uses AI
Both students and universities are struggling with how to incorporate AI into the higher education learning experience. Employers expect students to have AI skills when they enter the workforce, but that shouldn’t be an excuse for not considering how AI is used by students. Allowing them to use it just to find answers, or to outsource thinking isn’t preparing them for their future careers. Using it to augment their learning experience, so they develop the critical cognitive skills that are so crucial irrespective of whatever industry they end up in, will have a much greater positive impact.
