The AI turn: how research—and life—are being rewritten

With the magazine Time naming “The Architects of AI” as person of the year (1), what better moment to reflect on how AI is reshaping us as humans, and how this transformation echoes through the way we approach scientific work and understanding?

Don’t worry: we are not here to deliver another impersonal or unrelatable AI monologue; rather we hope to build, collectively, an honest reflection on a topic that is undeniably reshaping our lives. Even more than that, we would love this to be a conversation with you, as a way to stay awake to the world that is emerging around us.

Our everyday world is shifting: the way we work, the way we research, the way we understand ourselves as humans. And like many others we, as researchers, are beginning to ask how this is shaping us now, and what it might make of us later. Will our critical thinking soften at the edges, or will we simply move through ideas with a new speed—as if we have finally been handed the assistant that this fast-paced world always suggested we needed in order to be productive? In a nutshell, and with full honesty: it is hard to say. As researchers, though, we feel compelled to explore AI’s unpredictable implications with the same curiosity that has always driven us, and we’ll try to do so here once again by seeing both sides of the coin.

As the author of this blog post, I feel the urge to begin with the obvious irony: I’m writing about AI with the help of AI—and, perhaps, that is part of the lesson. As a non-native English speaker, I too make use of AI to refine my language and improve clarity, and I always try to do so with a grain of salt: I still want my thoughts to be mine and my voice to remain intact. Language editing, unsurprisingly, is in fact one of the most frequent reasons researchers turn to generative AI: a recent Nature survey found that, in the research community, the biggest impact of generative AI is to assist the writing of researchers who do not have English as their first language (2). This aligns with the open responses from our own survey we conducted on the potential impact of harmonizing grant applications (3): when we asked participants who use AI in their application process what their intended purpose was, “tweaking the language and improving clarity” emerged as the overwhelmingly common answer. Notably, the only participant who reported not using AI for grant writing was also the only native English speaker, and that contrast speaks for itself.

So, back to the topic of this piece, what sets our reflections apart is simple: we are trying to describe an unfolding transformation while we are actively participating in it, using the very tools we are questioning. Researchers usually observe systems from the outside at a comfortable distance, where trends can be studied with clarity; AI, though, for once collapses that distance. All of a sudden we are no longer observers of a system, but we are participants, immersed in the phenomenon itself even as we question it. This inevitably complicates how we analyze what’s happening.

Why AI is not just another tool

To call AI “just another technological innovation” is tempting, but also inaccurate. The calculator, the computer, and even the internet have transformed how we work and communicate despite initially facing some skepticism driven by the common and deeply human fear of the unfamiliar (4–6). Yet, none of these innovations has fundamentally challenged our sense of human agency. Although these tools expanded our abilities, we remained the ones deciding what to do, how to think, and which questions to ask. AI, on the other hand, seems to be working differently. As Yuval Noah Harari argues in his bestseller Nexus, AI is revolutionizing the way we deal with information, because it is information itself (7). By definition, information is “something that creates new realities by connecting different dots in a network”. In other words, it is anything that contains, carries, or transmits a message and helps us re-organize reality by establishing meaningful connections between its various parts. When something functions as information, it can unsettle the very identity structures we rely on to understand ourselves and our place in the world (8). That’s what AI does: it reshapes the conceptual environment in which the thinking itself occurs (9).

Ronald Purser, Professor of Management at San Francisco State University, makes this case provocatively in his blog piece titled “AI is destroying the university and learning itself.” (10) “Welcome to the death of higher education,” he writes—which may sound quite like a dramatic claim, yet it is one that echoes real concerns. Purser argues that while a tool helps us accomplish tasks, a technology reshapes the very environment in which those tasks, and the thinking behind them, unfold.

AI is different from the past innovations because it directly influences how we interpret reality, how we communicate, and even how we make decisions. Because it reshapes the human cognitive environment, it consequently and inevitably reshapes the scientific world, too; overlooking the broader societal and psychological effects of AI while discussing its use in research would hence be like looking at reality through a narrow lens and forgetting the broader picture that gives it meaning. This blog piece certainly cannot hope to capture the full complexity of this topic, but it aims nevertheless at touching upon a few essential elements that stir our curiosity and invite reflection.

When discovery speeds up: new frontiers of possibility

When we focus specifically on research advancements, AI advantages are impressive both in terms of quality and efficiency. A bibliometric analysis of research publications has demonstrated that AI has quickly moved from a computer-science niche to nearly every research field, with AI-related publications rising exponentially since 2015 (11) . This is because tasks that once required months can now be completed in hours and at a much higher level of complexity: the computing power used to train big AI models has been rising unbelievably quickly—ten times more each year, doubling every 3.4 months (12). AI can nowadays generate hypotheses, simulate scenarios, and uncover patterns that might otherwise remain invisible, which admittedly expands the frontiers of what is and will be scientifically possible. It can even assist in academic writing, with studies and reports showing that its use in research ranges from generating ideas to structuring content and editing manuscripts (13). The latter is, however, a slippery slope—one that requires particular caution, since generative AI is renowned to be prone to hallucinations (14) and its unethical use can compromise scientific integrity and good research standards.

As we dig into all this novel potential, we understand that the AI power is undeniably exciting and it points toward a future where scientific discovery reaches new depths, but it also reminds us that the future of discovery will depend on how wisely we use these new frontiers.

The fragile mind in a high-speed age

Where do we humans position ourselves in all these groundbreaking transformations? Is our mind still required to perform basic tasks and how are we affected by this technological outburst that already feels bigger than ourselves?

No one likes to think that we will be replaced by AI—and we can rest assured that this, most likely, will not happen. What is at risk though, or will at least require some level of adaptation, is how we are using our cognitive resources alongside these transformations. An increasing amount of studies has started to investigate the effect of large language models on human cognitive abilities as well as on our motivation in performing tasks, and has found that while AI increases productivity, it often reduces intrinsic motivation (15). Reliance on AI has been also associated with reduced memory retention, lowered analytical endurance, and a diminished capacity for problem-solving (16). To lean on methods that increase speed and productivity but reduce original thinking and creativity poses ethical questions. If a generation of researchers grows accustomed to outsourcing thinking in order to speed up productivity, what happens in the long run to the intellectual foundations of science? Perhaps, the enthusiasm surrounding speed in research deserves a second look. AI certainly helps us move faster. But who decided that faster is better? Who told us we have to be running? The academic pressure? The quiet cult of productivity?

From a more philosophical perspective, it might feel shallow to celebrate speed without asking whether acceleration leads to better thinking, or whether it simply compresses the time in which that thinking occurs. Humans do have a biological pace for processing thoughts (17) and perhaps it would sometimes be wiser to allow ourselves that time—to make meaningful discoveries happen, and most importantly to grow alongside the world we are building—rather than chasing an ever-growing number of findings whose real impact is yet to be determined. As the popular saying goes, after all, “Rome wasn’t built in a day”. Or is this still where we are aiming to go?

Toward tomorrow: a complex future, still written by us

AI is complex and surely powerful—but so are we, and so are research and scientific progress. Integrating it into the framework of a complex world while still making the most out of it is perhaps the real challenge of our days. As we touched upon in this piece, AI accelerates our work, but it can also erode motivation and cognition; it democratizes knowledge, but it can threaten accuracy; it expands what is possible, yet it challenges our assumptions about what it means to think, create, and understand.

For all the speculation about AI’s potential autonomy, one truth is often overlooked: AI is fundamentally shaped by human decisions. The datasets, the training methods, the objectives, the constraints—all of these reflect human priorities, human values, and human blind spots. AI is not an alien intelligence; it is a mirror where we merely see our aspirations, our contradictions, and even our vulnerabilities. This provides us with a unique opportunity: to deepen our appreciation for the complexity of human cognition. In teaching machines to imitate aspects of intelligence, we are forced to revisit our own understanding of what intelligence actually is. What does it mean to understand something deeply rather than merely reproduce or predict a pattern? What is the value of creativity that emerges from genuine curiosity rather than statistical prediction? What is uniquely human in the act of thinking? These and many more questions are still here for us to be explored.

In the meantime, one truth remains: for AI to be functional while still leaving us with the most human and most important thing we own—our sense of agency—we need to make a conscious effort. We still have the ability to set boundaries, ask questions, and protect the cognitive and ethical qualities that make research meaningful. Hence, care must be taken: not because AI is inherently dangerous, but because its power requires a thoughtful response. If the future is uncertain, we like to think that, as humans, we still have a choice in how we shape what comes next.

References

1. TIME Person of the Year 2025: How We Chose | TIME. https://time.com/7339621/person-of-the-year-2025-ai-architects-choice/.

2. Naddaf, M. How are researchers using AI? Survey reveals pros and cons for science. Nature https://doi.org/10.1038/d41586-025-00343-5 (2025) doi:10.1038/d41586-025-00343-5.

3. Gatto, F. Rethinking grant writing: can we do better? Researchers weigh in. AdvanSci Research Solutions https://www.advansci-research.com/post/rethinking-grant-writing-can-we-do-better-researchers-weigh-in (2025).

4. A Brief History of Calculators in the Classroom. Hack Education http://hackeducation.com/2015/03/12/calculators (2015).

5. Clive-Matthews, J. A brief history of tech skepticism. Strategy+business https://www.strategy-business.com/article/A-brief-history-of-tech-skepticism.

6. Trieste, L. & Turchetti, G. The nature, causes, and effects of skepticism on technology diffusion. Technol. Forecast. Soc. Change 208, 123663 (2024).

7. ‘NEXUS’ – A new book from Yuval Noah Harari (out September 2024). Yuval Noah Harari https://www.ynharari.com/book/nexus/.

8. Di Plinio, S. Panta Rh-AI: Assessing multifaceted AI threats on human agency and identity. Soc. Sci. Humanit. Open 11, 101434 (2025).

9. Abbosh, A., Al-Anbuky, A., Xue, F. & Mahmoud, S. S. Perspective on the Role of AI in Shaping Human Cognitive Development. Information 16, 1011 (2025).

10. Purser, R. AI is Destroying the University and Learning Itself. Current Affairs (2025).

11. Hajkowicz, S., Sanderson, C., Karimi, S., Bratanova, A. & Naughtin, C. Artificial intelligence adoption in the physical sciences, natural sciences, life sciences, social sciences and the arts and humanities: A bibliometric analysis of research publications from 1960-2021. Technol. Soc. 74, 102260 (2023).

12. Chojecki, P. AI Moore’s Law. Medium https://pchojecki.medium.com/ai-moores-law-18391003432e (2025).

13. Khalifa, M. & Albadawy, M. Using artificial intelligence in academic writing and research: An essential productivity tool. Comput. Methods Programs Biomed. Update 5, 100145 (2024).

14. Chelli, M. et al. Hallucination Rates and Reference Accuracy of ChatGPT and Bard for Systematic Reviews: Comparative Analysis. J. Med. Internet Res. 26, e53164 (2024).

15. Research: Gen AI Makes People More Productive—and Less Motivated. https://hbr.org/2025/05/research-gen-ai-makes-people-more-productive-and-less-motivated.

16. Kosmyna, N. et al. Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. Preprint at https://doi.org/10.48550/arXiv.2506.08872 (2025).

17. Nuwer, R. The Human Brain Operates at a Stunningly Slow Pace. Scientific American https://www.scientificamerican.com/article/the-human-brain-operates-at-a-stunningly-slow-pace/ (2025).LeBlanc, A. G., Barnes, J. D., Saunders, T. J., Tremblay, M. S. & Chaput, J.-P. Scientific sinkhole: The pernicious price of formatting. PLOS ONE 14, e0223116 (2019).

Author

Francesca Gatto is a PhD candidate in immuno-oncology at Karolinska Institutet, a former intern in Communication and Outreach and now a Guest Blogger at AdvanSci. Driven by a passion for meaningful dialogue between science and society, Francesca works to make research engaging, accessible, and inclusive for everyone.

Editor

Jane Fisher, PhD, is a co-founder of AdvanSci Research Solutions. She is passionate about practical solutions that enhance human health and improve the quality of biomedical research.