News
“We have to take control,” says philosopher Paulan Korenhof on AI
Developments in AI are progressing rapidly, and it seems as if no one can do without it anymore. Paulan Korenhof, who researches digital representations, raises critical questions about the way AI is being embraced.
Dr. Paulan Korenhof has a background in philosophy, law, and art, which gives her a unique perspective on how digital representations shape the world. “My PhD research focused on the representation of individuals, particularly the right to have personal information removed from the internet. At the Environmental Policy Group, I’ve been given the opportunity to study technologies that create digital representations of our environment—AI included.”
As an Early Career Scientist, Korenhof will also present her thoughts on AI during the Dies Natalis of Wageningen University & Research on March 7, 2025.
The need for transparency
Representations, such as ecosystem models, play a crucial role in environmental and policy discussions. A well-designed representation highlights key elements, making it easier to engage with complex issues. “In the future, self-learning algorithms and other forms of artificial intelligence will increasingly be used to generate these representations,” says Korenhof. “That needs to happen consciously.”
Models themselves already pose significant challenges. Korenhof quotes the Polish philosopher Alfred Korzybski, who said: “The map is not the territory.” In other words, a model should never be mistaken for reality—it is a selective interpretation. “If you let AI construct that model, you’re adding another layer between you and the world. Right now, experts still struggle to explain how artificial intelligence reaches its conclusions. That opacity could be a strong reason for caution.”
During my PhD, I worked at the Privacy and Identity Lab in Nijmegen. I was surrounded by hackers, which I found incredibly inspiring
The decision to use AI must be well-justified. “Climate change and biodiversity loss stem directly from how our world currently operates. If AI merely increases efficiency within our existing framework, we’ll remain stuck in outdated systems. Meanwhile, the scientific community is reaching a consensus that fundamental change is necessary.” Moreover, AI itself is still a polluting technology, requiring significant amounts of energy and cooling water.
How AI sees the world
“To some extent, artificial intelligence is a technology in search of a problem,” Korenhof argues. “There’s a kind of arms race to develop the best AI, but the approach isn’t problem-driven. Instead, people are searching for problems to apply AI to.” While AI undoubtedly has potential, Korenhof believes its applications are not always explored in the right direction. “The focus is on achieving human-like intelligence. It might be more interesting is to examine how AI uniquely perceives the world.”
One example comes from a study where researchers trained an image recognition algorithm to distinguish huskies from wolves. “The algorithm performed well. When researchers analyzed how it made its distinctions, though, they discovered that one of the key factors was the background of the image. If the background contained snow, the algorithm was more likely to classify the image as a wolf. There could be many ways in which AI can offer new perspectives on the world instead of mimicking our own.”
Korenhof sums up her stance: “As AI continues to evolve rapidly, we must collectively decide how to use this technology responsibly. We need to take control. If, fifty years from now, we look back and realize we’ve become more efficient but also more polluting, then nothing will have truly changed.”
Dies Natalis 2025: Artificial Intelligence for Sustainable Futures
On Friday 7 March 2025, Wageningen University & Research will celebrate its Dies Natalis. This year's theme is: Artificial Intelligence for Sustainable Futures.