MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) holds a day-long Agentic AI Summit on Nov. 9, according to its CSAIL Alliances event page. The program starts at 8 a.m. with registration and breakfast on the eighth floor of the Schwarzman College of Computing in Cambridge, Mass.
The lab describes agentic AI as systems that “move beyond generating content to taking autonomous, goal-directed actions in real time.” It says its principal investigators are working on topics including last-mile AI, security, deployment, cost and physical AI.
CSAIL director Daniela Rus opens the day, followed by a morning keynote from Dylan Hadfield-Menell, an associate professor in MIT’s electrical engineering and computer science department, according to the agenda. The lab notes that the agenda may change.
Most of the day is research talks from CSAIL faculty. Michael Cafarella, a principal research scientist, speaks on infrastructure and data. Daniel Jackson, a professor and CSAIL associate director, has a talk on using concept design, a method for structuring software, to make agents more dependable.
After lunch, Armando Solar-Lezama, CSAIL’s associate director and chief operating officer, gives a talk on safety and tools. Research scientist Erik Hemberg covers agents and cybersecurity, and professor Brian Williams of MIT’s aeronautics and astronautics department speaks on robotics and agents. Associate professor Jacob Andreas closes the research block with a talk on agent behavior.
Lunch is split into small-group topic tables with researchers, the agenda says. Hosts include Una-May O’Reilly, Manya Ghobadi, Hemberg and professor Wojciech Matusik. Ghobadi is also listed as founder and CEO of Systalyze.
Industry and safety panels
An afternoon panel on industry uses and open questions includes Hazel Dalton, a chief technology officer at Nasdaq, Amir Durrani of NTT DATA and Zoo Majzoub, a senior staff computer scientist at Adobe.
A second panel on agents and safety brings together Roger Roberts, a partner and co-chair of the McKinsey Technology Council, O’Reilly, Hadfield-Menell and Mok Oh, co-founder and CEO of Foundation EGI.
The last talk comes from Andrew Lo, a professor at MIT’s Sloan School of Management and faculty director of FinTechAI@CSAIL. It’s titled “AI Safety Is a Selection Problem: Alignment and Agentic AI from an Evolutionary Perspective.” The day ends at 5 p.m., followed by a reception.
Who can attend
The summit is for CSAIL Alliances members, the MIT community and companies interested in CSAIL research. CSAIL Alliances is the lab’s industry membership program. The lab says space is limited for outside attendees, who must apply and wait for a follow-up email.
In-person attendance is recommended, but there’s also a live stream, according to the event page. The page doesn’t list a price for outside attendees.
Source: MIT CSAIL Alliances, MIT CSAIL, “MIT CSAIL Agentic AI Summit”
