Meeting Register Page

Open Data Day: Designing for Data Rights in the AI Production Pipeline

Building trustworthy AI requires building public trust in how AI is developed. While the majority of AI production/resources are concentrated within a few companies in even fewer countries, alternative spaces are emerging for more people to participate in creating, applying, and governing data and data-generated ML models. New initiatives such as BigScience and BigCode seek to change extractive methods of AI production, replacing secretive web scraping with data stewardship and other data rights-affirming tools, practices, and systems.

Through a workshop conducted through Miro, we will walk through the current methods of data procurement for building ML models and discuss opportunities to integrate community preferences into the AI production pipeline. Target participants will include contributors to open source datasets or open source code repositories. The outcome will be a public Miro user flow diagram representing the diversity of perspectives from participants and summarising key takeaways from the discussion.

Mar 9, 2023 02:30 PM in London