According to Tornoreanu (2026), 92% of companies plan to invest more into AI in the coming years, but very few of these companies have a clear plan on where to spend that money. The companies don’t know how AI will affect them, but they still have a firm belief that it will. Somehow teachers are supposed to have a clearer vision, because the teachers are expected to teach how to function in a situation no-one understands.
Pathfinder (2026; LAB 2026) is a project at LAB in collaboration with partners around Europe. One of the challenges in a project like this is that the understanding around these tools is still in its infancy, especially regarding the long-term effects or how these tools are actually going to be used in working life.
The student
Only very few students really understand what artificial intelligence is. The usual understanding is limited to large-language models and other generative systems. With that in mind, students fall into three categories regarding their approach: some students don’t want to use AIs at all, some students see these as tools they need to learn to use, and some students just see AIs as plagiarism machines while still using them without much thought or planning.
The real problem here is that it is possible to produce assignments easily with LLMs, but even in the case of the relatively simple and controlled problems students are presented with, the results are generally not good. As Chasins puts it in her Tech Support (2025) video, LLMs are tools that were created to write good looking text, so they can write text that looks good. The subtext is that while this text looks good, it is not often actually meaningful. If you don’t have the underlying expertise to assess the text given by the LLM, using them is not useful. If you are reliant on these tools, you will never do the work required to gain this expertise. Also, the potential for hallucinations is high. This means that there is an element of luck which is hard to control.
It appears that many students are willing to sacrifice their grades for ease. Considering the pressures many of them are under, this is understandable. However, this produces its own pressures through ethical stress.
The teacher
For teachers, LLMs have made designing assignments more complicated. It is not enough to find an interesting problem based on the topic at hand. Now, the teacher needs to test their assignments with LLMs. At the same time, the teacher also needs to make sure the assignments don’t become overwhelming to those students who don’t want to use these tools.
What makes this even more complex is that LLMs are known to affect the cognitive capabilities of their users negatively (Liu et al. 2026; Bedard et al. 2026; Speri 2026) and people tend to believe LLMs even if they understand they hallucinate (Ipsos 2025). The long-term effects are still largely unknown. While attempts to fix these problems are under way, these fixes often require years of previous schooling around LLMs, which can potentially mean that these fixes will reach higher education in ten to fifteen years.
Why is this important?
Sam Altman has a vision that intelligence is a utility we pay for in the same way we pay for water and electricity (Okemwa 2026). This resulted in a widespread backlash. At the same time, many people are still using these tools and very few have paid attention to the reporting regarding their negative effects. At the same time, much of media, especially various self-imposed media bubbles, inundate people with the message that AIs are taking over everything.
LLMs work well when they support the user in their work, but the current tools have not been built from this perspective, instead they are designed to replace humans. This isn’t widely understood.
This leaves teachers in an awkward position. How to approach this pervasive phenomenon that wants to push workers out of the picture?
Author
Aki Vainio works as a senior lecturer of IT at LAB and takes part in various RDI projects in expert roles. His summer reading list for 2026 includes titles such as AI Snake Oil, Empire of AI and The War of Art.
References
Bedard, J. Kropp, M., Hsu, M. Karaman, O.T., Hawes, J. & Kellerman G.R. 2026. When Using AI Leads to “Brain Fry”. Harvard Business Review. Cited 7 May 2026. Available at https://hbr.org/2026/03/when-using-ai-leads-to-brain-fry
FRESCO. 2017. White, Wolves, Pack image. Pixabay. Cited 15 May 2026. Available at https://pixabay.com/photos/white-wolves-pack-nice-animals-2704666/
Ipsos. 2025. Audience Use and Perceptions of AI Assistants for News. BBC News. Cited 7 May 2026. Available at https://www.bbc.co.uk/aboutthebbc/documents/audience-use-and-perceptions-of-ai-assistants-for-news.pdf
LAB. 2026. Pioneering AI Technology in Higher Education to Facilitate Innovation and Nurture Development. Pathfinder. Project. Cited 7 May 2026. Available at https://lab.fi/fi/projekti/pioneering-ai-technology-higher-education-facilitate-innovation-and-nurture-development
Liu, G., Christian, B., Dumbalska, T., Bakker, M.A. & Dubey, R. 2026. AI Assistance Reduces Persistence and Hurts Independent Performance. Arxiv.org. Cited 7 May 2026. Available at https://arxiv.org/pdf/2604.04721
Okemwa, K. 2026. AI as a utility bill? Sam Altman thinks that’s the future. Yahoo! Finance. Cited 7 May 2026. Available at https://finance.yahoo.com/news/ai-utility-bill-sam-altman-180219388.html
PATHFINDER. 2026. Welcome to the Erasmus+ Pathfinder Project. Pioneering AI technology in higher education to facilitate innovation and nurture the development of entrepreneurial resources. Netlify. Cited 18 May 2026. Available at https://erasmus-pathfinder.netlify.app/
Speri, A. 2026. ‘I wish I could push ChatGPT off a cliff’: professors scramble to save critical thinking in an age of AI. Guardian. Cited 7 May 2026. Available at https://www.theguardian.com/technology/ng-interactive/2026/mar/10/ai-impact-professors-students-learning
Tech Support. 2025. Professor Answers Coding Questions (video). WIRED. Cited 7 May 2026. Available at https://www.youtube.com/watch?v=PZ_ebxkNZmo
Tornoreanu, V. 2026. The AI Productivity Paradox: When Efficiency Kills Demand. Cited 7 May 2026. Available at https://ssrn.com/abstract=6631838
