As part of NeurIPS 2020, PAI co-hosted a full day workshop on ‘Navigating the Broader Impacts of AI Research‘. The goal of the event was to encourage the research community to think critically about our roles and responsibilities in anticipating and mitigating potential negative consequences of AI research. It included panel discussions on impact statements, publication norms, harms of AI, and more. The event was part of our ongoing work on responsible publication norms.
We opted to have the majority of the workshop as panel sessions to encourage discussion on what we see as an important topic that requires community participation. There has been a lot of discussion on social media within the AI community on broader impact statements and ethical reflection, but the format is not always conducive to productive disagreement on what is a very nuanced issue. By hosting longer, structured discussions with a diverse set of panelists, we aimed to advance the conversation in an inclusive and productive way.
We were blown away by the thoughtful engagement of the panelists and the attendees, who were able to participate through commenting and asking questions via a text chat function.
We accepted 15 papers to the workshop, and as well as the panel discussions, we also hosted an interactive paper session on GatherTown. This allowed attendees to virtually wander around and speak to authors of the papers that had been accepted to the workshop.
I’d like to thank my co-organizers Carolyn Ashurst (Governance of AI, Future of Humanity Institute), Deb Raji (Mozilla), Solon Barocas (Microsoft), and Stuart Russell (Center for Human-Compatible AI, UC Berkeley), as well as all of our fantastic speakers, panelists, paper authors and reviewers, and all those who helped behind the scenes to make the day a success.
You can watch recordings of the panel sessions and the paper lightning talks below.
Recordings of live panel sessions and talks
Carolyn Ashurst (GovAI) and Rosie Campbell (PAI)