From our CLAI Opening Seminar
On Wednesday, September 20th, we hosted our opening seminar, marking the establishment of the Center for Language Generation and AI (CLAI) as a framework for text generation and AI research at Aarhus University. We couldn’t have asked for a better start. In a packed theatre hall at Aarhus University, we were fortunate to have a great lineup of speakers, who helped make the opening an important milestone for fostering future knowledge sharing and collaboration within this exciting field at Aarhus University and beyond.
In his opening talk, CLAI director Mads Rosendahl Thomsen gave us a sense of the rapid development in the field and the importance of importance CLAI as a center “without walls” that can connect scholars from various disciplines and backgrounds and foreground much needed Humanities research in the area.
In her presentation, CLAI co-director Rebekah Baglini instigated an insightful discussion on “grappling with a new cultural technology”, mentioning, among other things, the opaqueness of these technologies – also for “insider” researchers – and the current, increasingly hyped (media) discourse on the technologies. In this context, Baglini noted, CLAI can support the collaboration and knowledge sharing that can help reframe thinking about these novel technologies, towards “carefully designed, interdisciplinary, expert-led, non-commercial, human-focused research”.
Morten H. Christiansen's opening talk shed light on the transformative potential of generative artificial intelligence in the educational sector, highlighting not only the ethical and negative possible impacts of these technologies on education, but also the opportunities to rethink and reform. Christiansen also stressed that educators will benefit from AI but that it is essential to draw clear lines between what uses that are encouraged and which are prohibited.
Kristoffer Nielbo emphasized the importance of developing a Danish Foundational Model. With a team of young researchers the Center for Humanities Computing have already developed a prototype, which performs very well on certain task, for example classification. This also underlines that the development of LLMs is about much more than chatbots but the development of reliable resources that can be used in research in many fields.
Luciano Floridi shed light on the ethical complexities surrounding LLMs, addressing transparency, bias, and misuse concerns. Professor Floridi underlined the need for legislation to control the use of AI and urge humanities scholars to engage in this field.
As we move forward, we are filled with enthusiasm and a sense of purpose in facilitating knowledge sharing and collaboration within this field.
We also want to thank all participants for being a part of this significant milestone, and we eagerly anticipate the opportunities and challenges that await.