Creating guidelines on the safe use of Large Language Models
It’s undeniable how the assertive presence of generative artificial intelligence (AI) and Large Language Models (LLMs) has generated a number of extraordinary benefits, as well as risks for society and the scientific community. For instance, how the overuse of LLMs affects research quality and (scientific) researchers' responsibilities, or how tech companies have a greater impact on society at large. As a result of the rise of these concerns, a group of interdisciplinary researchers held a summit on June 19th , 2023, inviting representatives from numerous prestigious and pertinent organizations in Europe and the Netherlands.
Guidelines for a better future
The summit’s main goal was to develop new worldwide guidelines on the safe use of LLMs in scientific research, as outlined in Claudi Bockting and colleagues' Nature publication. Participants employed LLMs to construct "living guidelines" that would be sufficiently flexible to respond to opportunities and challenges brought on by the quick advances in AI. These "living guidelines" were developed in order to protect the quality and independence of scientific research, as well as provide (scientific) researchers the resources they need to fairly assess and certify LLMs in order to minimize any potential risks to both society and the scientific community.
The summit was organized by prof. Claudi Bockting and colleagues and hosted by the Institute for Advanced Study in Amsterdam.