Han van der Maas | The idea that STEM builds while the humanities and social sciences criticise is a cliché
Technological innovation does not come from technology alone: the social sciences and humanities can also have a major societal impact through AI, writes columnist Han van der Maas. He sees opportunities in areas including education, language and mental health.
This month, Peter Wennink, former CEO of ASML, visited the UvA to discuss what his report, “The Route to Future Prosperity”, means for the university. I was in the audience at REC Impact for a panel discussion on the role of universities in this route. This controversial report calls for more than €100 billion in investment in four strategic technology sectors. One concern is that my field, the Social Sciences & Humanities (SSH), will be left behind. The discussion did not go entirely as I had hoped.
When it comes to developments in AI, the contribution of SSH may appear to be primarily critical in nature. We raise ethical questions, examine the risks to democracy, call out the behaviour of tech giants, highlight the environmental damage, and so on. The AI, Media and Democracy Lab, for example, researches the responsible use of AI in the media.
But I am also somewhat critical of this contribution. First of all, taking a critical stance towards technology is not the preserve of any one academic field. The idea that STEM disciplines build things while the humanities and social sciences point out the problems is a cliché. Almost all AI researchers also have a nuanced view of the dangers of AI (see the Statement on AI Risk from 2023). Another good example is the Leiden Declaration on Artificial Intelligence and Mathematics, which is relevant to the recent news that OpenAI may have solved one of the Millennium Prize Problems in mathematics. Secondly, the contribution of SSH researchers is not merely critical in nature.
My own subfield within SSH, psychology, is in fact one of the foundations of modern AI. Many discoveries concerning neural networks and reinforcement learning, the two pillars of modern AI, originated in psychology. Geoffrey Hinton, one of the “godfathers” of AI, who was awarded the Nobel Prize in Physics in 2024 for his contributions to AI, studied experimental psychology. Many cognitive psychologists work at the intersection of brain research, cognition and AI, for example on visual perception and the ability to reason using analogies. This more fundamental research also regularly leads to the creation of spin-offs such as Oefenweb, an educational software company of which I was scientific director until a few years ago.
But there are also impressive examples of technological innovation in other SSH fields. A particularly notable example is the freely available phonetic software Praat, which has been developed at the UvA since 1992 and is now considered the standard in phonetic research worldwide (linguistics). The software is used in clinical speech and language therapy and by music professionals for voice and singing analysis. Another open-source contribution with applications around the world is JASP, a free alternative to SPSS. Yet another example is the digital tools developed by the Govcom.org Foundation (Media Studies), which are widely used in international research on social media. An example from the Faculty of Economics and Business is Veylinx/RevFixr.
I see major opportunities for the SSH field to contribute to technological innovation in the Netherlands, particularly by translating the latest developments in AI into new applications. This applies to education, media, language, mental health, and much more. Let me give an example. An estimated 10 to 30 percent of AI users seek help with mental health problems at some point. These are staggering numbers worldwide. This trend needs to be managed responsibly. A great deal of fundamental and applied research is needed for this, but so are the resources to build well-protected and regulated AI chatbots for mental health support, for which, as we know, there is an enormous shortage, with major societal costs as a result. In short, a responsible digital transformation, one of the UvA's new strategic themes, also involves building responsible new digital applications.
After the panel, I quickly walked to the front and had about 30 seconds to explain this to Peter Wennink. I failed. Peter quickly interrupted me and praised the psychology of leadership. That was the end of my turn.
