Artificial intelligence has developed quite a terrible reputation.
Depending on who you ask, it’s making students stupid, stealing artists’ work, taking our jobs, destroying the environment and possibly convincing us that the word banana has three R’s in it. Sure, some of those criticisms are fair, but they leave out half the story.
AI can be incredibly useful.
It can explain a concept five different ways until one finally clicks. AI can troubleshoot code, help you practise or learn another language, aid in preparing for a job interview, organize a chaotic schedule or figure out what to cook with the six increasingly questionable ingredients left in your fridge.
For students in particular, the possibilities are obvious.
AI can generate practice questions for exams, walk you through difficult concepts, give feedback on writing, or help turn a vague idea into a starting point. For some students with learning disabilities or neurodivergence, it can break overwhelming assignments into manageable steps, reorganize information, or translate scattered thoughts into something they can actually work with. Personally, I have seen firsthand how useful this can be. Through the use of platforms such as Anki, Perplexity AI or Notebook LM, these make an already challenging learning environment for those who are neurodivergent and/or have learning disabilities much easier to navigate.
In this case, AI isn’t always simply doing your thinking for you; oftentimes it is helping you think. I find that this distinction gets lost in conversations about AI use in university, when every assertion seems to begin and end with cheating.
Obviously, asking ChatGPT to write your paper and then submitting its response isn’t learning. However, asking it to explain why your statistical analysis is wrong is not the same thing. Neither is asking for quiz questions before an exam.
Could we become intellectually lazy if we outsource everything to AI? Absolutely. But that is an argument for how to use it, not against using it at all.
Think of the implementations of other technologies. A calculator, for example, is useful because we still understand what numbers mean and the math behind the calculations. Google is useful because we can evaluate what we find. AI should be treated the same way: as a tool whose usefulness depends largely on the person using it.
This usefulness includes research.
AI can help researchers brainstorm ideas to fuel their research, organize information, identify patterns in large datasets that may have otherwise been missed for example by using AIRE by Briya (a healthcare and medical research platform), and automate repetitive tasks that would otherwise take hours using Causaly for example. It can also make complicated research more understandable to people outside academia. AI aids in academic accessibility.
None of this replaces the researcher. Someone still needs to understand the methods, verify the information, interpret the results, and recognize when AI is confidently talking nonsense.
But this is the version of AI that is worth pursuing. Not replacing human expertise, but in support of human endeavors. Often, AI can give humans the gift of more time after use.
And that potential extends beyond university!
As previously mentioned, AI can translate languages, make information more accessible, and take over tedious administrative work. A health-care professional could spend significantly less time on paperwork and more time actually with patients. A small-business owner could automate work they cannot afford to hire someone else to do and even help someone staring hopelessly at a government form by asking AI to explain the form using different language or phrasing.
Not every task becomes more meaningful simply because a human is forced to do it.
And I know what those who are strictly anti-AI will say: what about artists whose work is used to train these systems? Or the people who are turning to chatbots for advice instead of therapists? What about jobs disappearing or the environmental cost?
Those concerns are real. But being pro-AI does not mean ignoring those issues. It means recognizing that the answer to harmful uses of AI is not necessarily less AI. It is better AI, with better rules around it.
Creators should have control over how their work is used, with protections around consent, attribution, and compensation. AI can help people talk through everyday problems, but it should not become an echo chamber or substitute for professional help in a crisis. If AI displaces workers, governments should invest in retraining and transition supports rather than leaving workers to absorb the cost.
And because “the cloud” is, unfortunately, not actually a cloud, companies profiting from AI should be responsible for reducing the energy, water, and physical resources their infrastructure consumes.
The basic principle should be simple: the greater the potential harm, the stronger the safeguards.
Universities should take the same approach, don’t simply ban AI or leave the responsibility to individual professors to do as they see fit. Teach students how to use it, how it fails, when not to use it and how to verify what it tells us. Then, design assessments that make us demonstrate what we actually know.
But one thing the anti-AI movement has definitively got right is that we should be skeptical of this new technology. We should question where its training material comes from, what biases it reproduces, what resources it consumes and who benefits from it.
But skepticism is not rejection.
AI can make education more accessible, accelerate research, remove tedious work and help people learn, communicate, create, and problem-solve. However, it can also help people practise academic dishonesty, spread misinformation, exploit creators, displace workers or reinforce someone’s worst ideas.
Both sides do hold truth.
But, ultimately, AI isn’t going anywhere. We can spend the next decade trying to put it back in the box, or we can build the rules, safeguards and literacy needed to make it work for us. AI is inevitable, but what we do with it isn’t.
