How do researchers feel about the potential impacts of AI on research?

We've been hearing about the dangers of AI for some time, but this month, the conversation turned existential. Warnings from the very people building these systems have some doubting whether it's more hype strategy than genuine alarm. But for a lot of people, the fear is real.

Academia is no exception. Researchers are, by definition, at the forefront of knowledge production — the very thing AI aims to do, eventually, better than humans. As it takes on more of our tasks, from the mundane to the more complex, what does science stand to gain, and what do researchers stand to lose?

AI is already pushing science toward breakthroughs once thought decades away. But the threats aren't only the worst-case scenario kind; there are quieter questions too, about what happens to human thinking and creativity when AI starts doing the thinking for us, even the easy parts.

We asked some of our senior ICTA-UAB researchers three questions: how worried they are about AI, what risk concerns them most in their own field, and whether AI has genuinely advanced their own work. Here's what they told us.

 

Eric Galbraith — Earth Systems

"I'm very worried. It is quickly becoming too powerful, and nobody is in control."

"I think the greatest risk is that we let the machine do our thinking for us, and we become dumber as a result. Learning takes effort, and if we offload to AI the effortful tasks that produce learning — reading, writing and arithmetic — we could diminish human capabilities. But it doesn't need to be this way. I think we could potentially have AI that aims to enhance human intelligence, rather than to replace it."

On his own use: "I use it to bounce ideas off, to explain things I am not familiar with, to summarize existing literature, and as a critical reviewer of things I've written. I am interested in working on AI as a means for improved democratic decision-making, but haven't done much in this direction."

 

Graham Mortyn — Paleoceanography

"I feel moderately concerned. We've seen hype over comparable issues in recent times."

"Excitement over the power and utility of this great AI tool must always be tempered with common sense and ethical considerations. This is essentially true of any technology that comes along, in my opinion."

He points to a specific shift in his own field: "I have seen written levels of English significantly improve in recent times on the part of non-native students. I do sensibly suspect AI is being tapped for improving writing, and this shows also in research. There is nothing wrong with this per se, though we do need to safeguard against misuse and plagiarism potentially. How exactly, I am not sure."

Where it's helped: "I particularly like PerplexityAI as a simple tool to ask basic questions and quickly derive a sensible answer. What I like about it is that it forces the user to ask clear, answerable questions — otherwise it spins its wheels in confusion. I have invoked some of this in my teaching in recent times."

 

Panagiota Kotsila — Climate Justice

"I'm worried that there is too much attention on 'what AI can do' and less on what the people behind it are allowed to do."

"The majority of people do not quite understand how AI works, the resources needed to support it, and the social risks involved. This worries me because ignorance, misunderstanding and low regulation offer the perfect breeding ground for manipulation and mistakes."

She's also seen this play out directly in teaching: "The most obvious impact I've seen at work is too many AI-generated papers, emails and essays from students. This creates an impression of better performance, but at the same time it's reduced our thinking and writing capacities — recycling ideas and thoughts in the most boring way. I think it can be detrimental for learning at school and at an academic level for younger generations."

On her own experience: "I feel it has only saved me a few minutes' worth of mental effort or simple work. I think it can be more useful for people working with big data. I've used it to summarize notes, but I never trust it much, so I've had to review it, which also takes time. I've used it for transcriptions and translations, and there it has been useful in terms of speed — although a person would have done a better job without needing to review it."

 

Álvaro Fernández-Llamazares — Ethnobiology

"AI is opening new possibilities, but it also raises important questions for the field of ethnobiology."

"For the Indigenous Peoples with whom I work, there is a real risk of reproducing old forms of knowledge extraction in a new technological format. Ethnobiology has generated huge amounts of knowledge about Indigenous Peoples' relationships with plants, animals and landscapes. As this knowledge is used to train AI, we need to ask: who has consented, who controls the data, and who decides how it is used? AI might recognise that a plant is medicinal, but miss the language, people, place, relationships and responsibilities that give that knowledge its meaning.

"Behind every AI system are data centres requiring enormous amounts of electricity, water, minerals and land. In the US, the expansion of data centres on or around Tribal lands is already raising questions about environmental justice and land-use decisions.

"For ethnobiologists, these are two sides of the same story: the extraction of knowledge and the extraction of resources. The challenge is to ensure that AI does not become simply the latest frontier of extraction, but a technology that respects Indigenous knowledge, data, lands and sovereignty."

 

Giorgos Kallis — Degrowth

"I am worried about the impacts it can have on social relations, the environment, and on redistributing wealth upwards."

"As young people use AI more and more, I fear that the basic human skills necessary for research will go extinct and eventually render human researchers useless. What skills should we teach young researchers today so that they remain necessary in the future? I feel that teaching research as if nothing has changed is irresponsible, but I'm not sure what to do differently."

On his own use: "I've used it for complementary tasks, like helping me with the tone of an email or for feedback on the weaknesses of a paper, that I then work to improve. But it always leaves me with a bad aftertaste, as if I ate something very sugary. Even with these small tasks, I am worried I am eroding some of my own skills and becoming one notch more stupid."

 

Isabelle Anguelovski — Urban Environmental Justice

"I'm concerned that AI will progressively take over our decisions and our capacity to think, where we're left simply giving instructions — if that."

"In my field the risk of replacement is lower, because interpreting and theorizing aren't easily automated. In social sciences, urban planning and political science, where I do see risk is in research support tasks — transcription, translation, graphic design — more than in the intellectual core of the work."

On her own use: "I use it mostly in heavy administrative tasks that aren't part of the core of my work. It helps me organize expenses and draft financial reports. It frees up time for research and reflection."

 

Arnim Scheidel — Political Ecology

"It's worrying that so few tech magnates and governments control AI, given their conflict of interest in meaningful regulation."

"AI systems require vast amounts of water, energy and minerals. Where will these resources come from? From whose lands?"

On his own use, he points to a concrete example from his research: "In a pilot project, we have used LLMs to code 30 years of news coverage on opposition to renewable energy projects in Spain. The set-up and quality control took a while. Once done, the LLMs coded thousands of newspaper articles with high-quality results within 2 hours. Manually, this would have taken probably a year of work of one person fully dedicated to it."

 

Patrizia Ziveri — Oceanography

"What worries me most is its unethical use, because AI is much more than just a new technology."

"One risk I see is that research could become more alike, with less room for creativity and real discovery. There's also the open question of intellectual property. But AI can be a great help too — during expeditions, for example, it can help us solve problems quickly, and it makes it easier to build on earlier work.

"I've used it mainly to save time on tasks around my research rather than on the science itself. I use it for conceptualisation, visualisation, polishing  documents and coding support. This leaves more time for thinking, interpreting data, and discussing and developing new ideas."

 

André Colonese — Early Food Systems

"The speed at which things are moving is alarming. The Industrial Revolution changed our way of living on this planet, and here we are talking about a completely different level."

"This is a clear turning point for our societies and possibly even for our species. AI offers many advantages but entails a potential future loss and reorganization of human skills as we become increasingly dependent on its benefits. The challenge here is to try to anticipate these changes."

On his own use: "I use it to help with coding. Something that used to take me days now takes just a few minutes or hours to get done.”

What emerges from these reflections is a certain tension between pragmatism and prudence, a recognition of what AI can improve alongside a real unease about what it might take from us.