AI shaming is not AI literacy

I have had enough conversations about generative AI (GenAI) to know that people have strong feelings about it. Some are excited. Some are tired of the hype. Some are worried about assessment, authorship, bias, copyright, over-reliance and academic integrity. As an educator in computer science, I share many of those concerns.

A recent observation highlighted negative attitudes: public messages that described GenAI users as “losers” or implied that people who use these tools are stupid. I do not think the intention was necessarily to harm anyone. It may have been meant as humour or simply to express frustration. But as an educator, it made me pause. 

It wasn’t the criticism of GenAI that unsettled me. It was how quickly criticism of technology turned into ridicule of people.

GenAI deserves serious critique. It raises important questions for educators and university leaders. But AI shaming is not the same as building AI literacy. In fact, it could hamper efforts to encourage responsible AI use. The language we use around AI can shape whether students and staff ask questions, disclose use or hide what they are doing.

The problem with shaming is that it treats AI use as one behaviour. It is not. A student might use GenAI to produce work they do not understand and submit it as their own. That is a clear academic integrity problem. But another student might use the same technology to understand a difficult concept, check the clarity of a paragraph, generate practice questions, plan a structure, translate their own ideas into more fluent English, test code or overcome the fear of starting.

• Spotlight guide: Boosting AI literacy across your institution
• An insider’s guide to how students use GenAI tools
• Why AI adoption in universities is really a question of trust

Staff use is varied too. A colleague might use AI to draft an agenda, simplify administrative text, generate teaching examples, prepare explanations, organise notes or make communication clearer. Even when AI is not being used as a formal adjustment, it might still help people manage cognitive load and communicate their thinking more clearly.

I use GenAI in my own work but not as a replacement for judgement. The ideas, priorities and responsibility remain mine. The tool helps me organise a messy starting point, test wording and make my thinking clearer. That experience has made me more careful about making assumptions about others’ AI use.

If we want students and staff to use the tools responsibly, we need to move from judgement to diagnosis. Instead of asking “did they use AI?” as if that settles everything, we need to ask better questions. What was the tool used for? Was that use allowed? Was it disclosed? Did the student check the output? What did they change? What human judgement remains? Can they explain the work? Does the submission still evidence the intended learning outcomes?

These questions are more useful than slogans because they lead to teaching, assessment design and fairer decisions.

There is also a trust issue. Many universities now ask students to declare or explain their use of GenAI. That is sensible, but disclosure depends on the surrounding culture. Craig Gonsalves’ work on AI-use declarations at King’s Business School found substantial non-compliance, with 74 per cent of students failing to declare AI use despite a declaration being required. That does not mean students are automatically dishonest. It does mean we cannot assume disclosure will work if the culture around AI use feels judgemental or unclear. 

Shame does not protect academic integrity. It often pushes the behaviour we most need to discuss underground.

For educators, the question is how to create a culture where students can talk honestly about AI use while still understanding that some uses are inappropriate. I think this starts with three shifts.

First, distinguish more clearly between substitution and support. If a student uses AI to substitute their thinking, authorship or evidence of learning, that is a serious concern. If they use it to support planning, revision, accessibility or formative understanding, the issue becomes more nuanced. Guidance should make those distinctions visible.

Second, teach verification as a core academic skill. Students should not be told simply that AI is useful or dangerous. They should be taught how to check claims, identify fabricated references, compare outputs, protect private data, recognise bias and decide when AI is unsuitable.

Third, design assessments that make human judgement visible. In an AI-rich environment, polished final text tells us less than it used to. We might need more emphasis on process, explanation, critique, version history, oral defence, reflective commentary, annotated outputs or task-specific justification. The aim is not to make every assessment longer. It is to make the evidence of learning clearer.

Language matters in all of this. Instead of “AI users are lazy”, we can ask what part of the work AI supported. Instead of “AI work is worthless”, we can ask what the student verified, changed and owned. Instead of “students who use AI cannot think”, we can ask whether the student can explain and defend the work.

This is not about lowering standards. It is about making standards teachable.

GenAI deserves serious critique. There are times when its use is inappropriate. There are tasks where it should be restricted. There are risks that students and staff need to understand properly. But ridicule is not a policy. Shame is not a pedagogy. Contempt is not a substitute for assessment design.

AI literacy should help people make better decisions. It should not make them afraid to admit that they are using tools already shaping education, work and society.

Sara Saravi is deputy lead of the Centre for AI in Education at Loughborough University.

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.