Dartmouth AI at 70 Conference Opens with Yann LeCun Keynote

Dartmouth College's AI at 70 conference opened October 29 in Hanover, New Hampshire, bringing together researchers, ethicists, and industry leaders to examine where human judgment must remain essential as artificial intelligence systems scale. The three-day event marks seven decades since the Rockefeller Foundation funded the 1956 Dartmouth workshop that formally named the field of artificial intelligence and established its foundational research agenda.

Yann LeCun, Meta's Chief AI Scientist and 2018 Turing Award winner, delivered the opening keynote address. The program centers on a single question: what aspects of human decision-making, creativity, and judgment should remain under human control even as AI capability expands across domains previously considered uniquely human.

Conference Structure and Format

The event officially launched on October 29, 2026, spanning three days of keynotes, panel discussions, and short-form talks designed for cross-disciplinary dialogue. Organizers deliberately assembled speakers from different fields who hold divergent views on AI's value, costs, and limitations. Sessions are structured to encourage debate rather than consensus, reflecting the field's current uncertainties about its own trajectory.

All sessions are livestreamed free of charge on YouTube, with in-person attendance still available for those who can reach the Hanover campus. The program runs on Eastern time, with session recordings accessible to international audiences who cannot attend live.

Student Hackathon Component

An Anthropic-sponsored student hackathon runs parallel to the main conference, culminating in final presentations and awards at the closing session. The hackathon gives undergraduate and graduate students hands-on experience building AI systems while engaging with the conference's central ethical questions about human agency and AI automation.

Student teams present their projects to a panel of judges drawn from academia and industry. The hackathon structure mirrors the original 1956 workshop's emphasis on practical experimentation alongside theoretical discussion, though participants now work with neural networks and large language models rather than the symbolic systems that dominated early AI research.

Historical Context and Founding

The conference commemorates the summer 1956 Dartmouth workshop organized by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon. That gathering, funded by the Rockefeller Foundation, produced the term 'artificial intelligence' and outlined research programs that shaped the field for decades. The original workshop proposal expressed confidence that 'every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.'

Seventy years later, the AI at 70 conference revisits that founding optimism with a more cautious lens. Where the 1956 participants focused on what machines could do, the 2026 program asks what machines should not do, or where human oversight remains non-negotiable even when automation becomes technically feasible.

Release Date and Availability

The conference officially launched on October 29, 2026, with sessions continuing through October 31. Registration for in-person attendance remains open, though organizers recommend checking the conference website for room capacity updates. All livestreamed sessions are free and require no advance registration, making the content accessible to global audiences interested in AI ethics, policy, and technical development.

What This Means

Dartmouth's AI at 70 conference arrives at a moment when artificial intelligence systems demonstrate capabilities the 1956 founders could not have anticipated, from generating photorealistic images to passing professional exams. The event's framing, 'What Must Remain Human?', signals a field grappling with its own success. As AI systems automate tasks once thought to require human intuition, the question shifts from technical feasibility to normative boundaries. The conference's diverse speaker roster and open debate format reflect a maturing discipline willing to question its own trajectory rather than assume all automation represents progress.