UC Berkeley’s second Agentic AI Summit drew thousands of students, tech professionals and researchers Aug. 1-2 for a weekend conference centered around the new evolution of AI that can act on its own.
Dawn Song, co-director of the Berkeley Center for Responsible Decentralized Intelligence, which hosted the event researches AI safety and responsibility — something she said was an important conversation at the summit.
“Agentic AI has the potential to transform how we work, learn, and solve complex problems, but realizing that potential requires us to think carefully about how we build and deploy these systems,” Song said in an email. “What has been encouraging to see at this summit is the willingness of people across the ecosystem to engage with both sides of that challenge.”
The event brought together about 5,000 attendees from more than 250 schools and 1,500 industry organizations, with 100,000 people watching the livestream, according to organizers.
The conference featured dozens of panels and presentations from more than 200 speakers — including UC Berkeley Chancellor Rich Lyons and Amazon, Nvidia and OpenAI executives and scientists.
“We’re a huge public university that really spans the range of knowledge. We’ve got departments in anything you could possibly imagine,” said Eric Meyer, dean of UC Berkeley’s School of Information. “This is the right kind of place for people to be talking about what the frontiers look like in AI — both the promises and the perils of that.”
For UC Berkeley junior Arjun Krishnan, there is “uncertainty” around how AI will evolve. As an environmental economics and policy major, he says AI also has the potential to solve human problems.
“Both sides of the coin can use it, and I think while we see a lot of corporations going kind of gung ho about using these technologies for ‘evil things,’ you kind of have to fight fire with fire,” Krishnan said.
Dozens of AI researchers also lined the halls of the Martin Luther King Jr. Student Union Building to share research spanning from AI regulation to its applications in science.
Jiaxin Pei, an assistant professor at the University of Texas at Austin and a researcher at Stanford, studied how AI systems had “hidden agendas”: system prompts that could override the user and secretly change the behavior of AI systems.
“Together we must — and I’m sure that we can — honor the ultimate objective of this summit by taking steps for a safe navigation towards a technology future that doesn’t just automate actions, but actively empowers individuals, enriches societies and amplifies human potential,” Lyons said during his opening speech. “There may be no higher calling at this moment of time in world history.”