Artificial intelligence has truly burst into our lives to such a degree that you cannot avoid it, in some instances, even if you wanted to turn it off. For example, when you search on Google, it now displays an AI-generated response at the top, whether you want it or not. Is it always correct? Some may assume so, but that is dangerous!

It has been proven time and time again that artificial intelligence isn’t always correct, sometimes it downright lies to your face or makes up things that aren’t real. This has several names: hallucinations, delusions and confabulations.

IBM describes AI hallucinations as follows: “AI hallucination is a phenomenon where, in a large language model (LLM), often a generative AI chatbot or computer vision tool, perceives patterns or objects that are non-existent or imperceptible to human observers, creating outputs that are nonsensical or altogether inaccurate.” (IBM Website, 2025)

This becomes a problem when conducting research, if one is using artificial intelligence to find resources and it hallucinates authors, journals or information. If too much trust is put into AI and the individual does not think critically of it or just copy paste without checking if the sources are true. They will end up using false or non-existent material as a base for their research.  This approach chips away at the integrity of research and academia. It becomes more difficult to ensure, that what is stated is correct, especially if no one checks whether the text cited supports the text and topic.

What is academic and research integrity really and why is it important?

Integrity is about being honest, trustworthy and truthful in one’s work, as well as acknowledging those who have added substance to one’s work and research methods. It cements, that what is being researched has substantial evidence either for or against your objective, in connection with research conducted by others. It shows that it isn’t just something you thought about on a Wednesday afternoon while drinking tea, and therefore it must be so.

Elementary School Critical Thinking

When I was in elementary school and we had to write essays about world events, we had to learn how to be critical of our sources, which news papers and other outputs could be trusted for a neutral, factual view of the topic. In order to gather as much trusted information as possible on the topic to properly analyse and understand the topic and then write an essay about it.

First we found the resources and then we began writing. I use the same formula when I write my blog posts, though a lot of what I write are my own thoughts on the topics. But first I read about the topic, from various trusted journals, news outlets, essays, conference papers etc.. to gain a deeper understanding for the topic, and to discover what other people have found through their empiric or theoretical research. Even if it defies my own initial theory, I gain a better understanding of the topic from different perspectives other than my own opinions and irrespective of my own biases.

So how can we use AI as a tool to reinforce integrity in research and academic writing?

AI is a tool and as we have discovered AI does make mistakes! If you use it to gather sources and citations, check every single source yourself. Read the text through, learn something from the text and ensure that it says, what you think it says. This way your integrity is protected and you will be sure, that what you reference supports your writing and research.

I’m not saying, that AI can’t be used, but we must be vigilant and smart about how we use it. Trusting AI blindly may put you or your organisation in trouble in terms of legality and integrity.

Use AI with human intelligence.

IBM. (2025, April 11). What are AI hallucinations? IBM. https://www.ibm.com/think/topics/ai-hallucinations