AI & Humanities: ‘So much untapped potential’
The field of artificial intelligence has developed rapidly in recent years. We spoke with Stephan Raaijmakers, professor by special appointment in Communicative AI, about the impact of artificial intelligence and why everyone should pay more attention to developments in this field.
The AI & Humanities, Help, Hype or Hassle event, on Thursday 7 April, aims to get the message of AI across to the humanities, according to Raaijmakers. ‘We want to show how beneficial an interaction between the humanities and the field of artificial intelligence could be. It’s an invitation to everyone in the humanities to open the discussion and see what each of us has to offer, but anyone with an interest in AI is welcome.’
Omnipresent but invisible
But why should artificial intelligence be of interest to the average person? ‘Artificial intelligence is a lot like electricity: it’s everywhere around us but we don’t see it. Take Google Translate, for instance. It contains language models that we know contain all kinds of prejudices,’ Raaijmakers explains. ‘If these models are based on data collected from a country with a selection bias, it will be reflected in the translations. Take gender bias as an example. If it is uncommon for men to work as a nurse in certain countries, this bias will creep into the system through the data. That’s why, when you translate a text, the output can look quite strange.’
This example may seem innocent, but it is also possible to suffer more directly from biases in the system, whether you are aware of them or not. ‘Suppose you are looking for a job and you upload your CV to a database. If the algorithm checks CVs for matches in a specific way and your CV doesn’t meet those requirements, it may end up falling between the cracks. You are basically at the mercy of the built-in prejudice of the system,’ he explains.
Opportunities for the humanities
AI can be an invisible obstacle in daily practice, and this offers opportunities for the humanities. ‘Ethics and morality are huge topics of research. How can you teach machines and algorithms human ethics and morals? People are also inconsistent in this respect of course. There are all kinds of issues where the humanities can get involved,’ says Raaijmakers.
‘And conversely, what kind of role could artificial intelligence play in research in the humanities? I think there is so much untapped potential.’ AI is currently generally used to analyse and measure data, but Raaijmakers sees opportunities to use the technology in different ways as well. ‘Take my field of Communicative AI, for instance. We used to only consider the ways people communicate with other people, but there are now a lot more datasets of interactions between people and chatbots. A lot can, and does, go wrong during these interactions. This is fascinating, because it allows you to analyse the ways in which people communicate with machines compared to how they communicate with people.’
Help, Hype or Hassle
Raaijmakers is not just excited about the future of AI in his own field. ‘The Humanities has all kinds of fields, like Art History, for example, where artificial intelligence could play a significant role.’
At the same time, he would like to temper some of the ideas about the fast rise of AI. ‘Many people assume that AI will take over the world in no time, but this is absolutely not the case,’ says Raaijmakers. ‘However, it is important to build and maintain communicative relationships with these systems, to make sure people stay in play. This is a hassle, as artificial intelligence is becoming increasingly technical. New developments are moving fast and you have to keep up.’
Interested in learning more about artificial intelligence? Join us at the AI & Humanities, Hassle, Hype or Help symposium. During the symposium, researchers and artists who work closely with artificial intelligence will discuss what AI means for the humanities, and vice versa. After the discussion there will be a screening of ‘Ich bin dein Mensch’. You can sign up for the symposium here.