Abstract
The release into the public domain and massive growth in the user base of artificial intelligence (AI) foundation models for text, images, and audio is fuelling debate about the risks they pose to work, education, scientific research, and democracy, as well as their potential negative impacts on cultural diversity and cross-cultural interactions, among other areas. Foundation models are AI systems that are characterized by the use of very large machine learning models trained on massive unlabelled data sets using considerable compute resources. Examples include large language models (LLMs) such as the GPT series and Bard, and image generator tools such as DALL·E 2 and Stable Diffusion. This discussion paper focuses on a widely used foundation model, ChatGPT, as a case study, but many of the points below are applicable to other LLMs and foundation models more broadly. UNESCO Catno: 0000385629
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
