In Our Own Image? SuperSalon with Irene Solaiman Exploring Ethics and Values in AI

In this SuperSalon, AI & technology safety expert Irene Solaiman (ex-OpenAI, currently at Hugging Face) will spark discussion around AI & ethics, values and natural language processing, and fairness and algorithms. The SuperSalon is hosted by ii member and digital wellbeing researcher Nastasia Griffoen (GEMH Lab | University of Twente) together with AI artist and researcher Mirabelle Jones (University of Copenhagen). 

Irene Solaiman is an AI expert, specialising in issues of social impact and international security. She formerly led public policy at OpenAI, where she initiated and led social impact and fairness research.

In this SuperSalon, Solaiman will give a short talk about her research in which OpenAI’s GPT-3 language model was fine-tuned on a value-sensitive dataset to produce favourable results. Following the presentation, discussion will be opened to all participants to explore issues of algorithmic (in)justice, values-sensitive natural language processing, and ethics in AI.

Recommended reading includes:

• the PALMS paper,
• Release Strategies and Social Impacts of Language Models,
• GPT-3 Its Nature, Scope, Limits and Consequences,
• On the Dangers of Stochastic Parrots,
• An Artificial Intelligence Helped Write This Play. It May Contain Racism.

Are you intrigued and interested in getting some hands-on experience with embedding your values into AI? Check out this free companion-workshop organised by the hosts on May 26th, Turn Yourself Into a GPT-3 Chatbot Using Social Media!

***

🗓 ii Calendar

📋 Code of Conduct

📚 Become a member, get a free ticket every month (FAQ), access our forums, members-only events, and more!