NEWS How can researchers in the Humanities meet the challenges posed by ChatGPT and other applications of generative AI? This was the theme of Huminfra's first webinar for which over 100 participants across Sweden registered.
Professor at Humlab
ImageHans Karlsson"We are very happy that so many people wanted to participate. It shows that the work done within Huminfra is important, which was also confirmed when the Swedish Research Council granted extended funding. Humlab at Umeå University has been a driving force in the development of digital humanities in Sweden and it is gratifying that the area is strengthened through Huminfra as a national infrastructure", says Coppélie Cocq, deputy director of Huminfra and assistant director of Humlab at Umeå University.
The webinar began with a keynote by Professor Geoffrey Rockwell, University of Alberta, Canada, who gave a historical perspective on how the ability to conduct a dialogue has dominated what we humans include in the concept of intelligence. He also discussed the development of AI, which took a big step forward in the 1960s when Joseph Weizenbaum created Eliza - one of the first chatbots. Eliza mimicked users' conversation patterns, creating the illusion that the system had human characteristics and contributing to the coining of the term The Eliza Effect.
We know that the systems are not human, but they are so good at imitating us that we talk about them in terms we would use about a human.
“Intelligence can deceive through imitation. We know that the systems are not human, but they are so good at imitating us that we talk about them in terms we would use about a human. This means we attribute characteristics to them that are not justified because they are words put together by computers.”
In a subsequent panel discussion moderated by Mats Fridlund, part of Huminfra’s leadership and researcher and Deputy Director of the Gothenburg Research Infrastructure for Digital Humanities (GRIDH), four researchers gave their perspectives on how research in the Humanities contributes to the development of AI.
"We have lacked a debate in Sweden about the challenges of generative AI for humanistic research and knowledge production, so this conversation became an inspiring starting point for this discussion. This was the first but by no means the last word on the matter”, says Mats Fridlund.
Generative AI such as ChatGPT is designed polite to encourage us to use the tool.
One of the panelists was Karin Danielsson, director of Humlab at Umeå University and docent in Informatics. One of her research interests is AI and user participation.
Karin Danielsson, director at Humlab
ImagePer Melander"Generative AI such as ChatGPT is designed polite to encourage us to use the tool. How and with what we train the systems in the time coming affects what kind of society we will have in the future and here there is much that research in the humanities can contribute to, such as analyzing in comparison with artificial intelligence, what are the behaviors, characteristics and abilities we value in humans?
Huminfra is a Swedish national infrastructure supporting digital and experimental research in the Humanities by providing users with a single entry point for finding existing Swedish materials and research tools, as well as developing national method courses. Huminfra is a consortium consisting of 12 nodes across 11 universities and organisations, coordinated by Lund University Humanities Lab.
huminfra.se
