These provide an opportunity to engage deeply with seminal and/or recent AI safety work.
In the past we've covered both technical research and AI governance/policy/law papers.
Anyone can suggest and briefly present a paper, before the group discusses it together.