Dartmouth opens ‘AI at 70’ conference with human judgment at the center

Dartmouth opened its “AI at 70: What Must Remain Human?” conference on October 29, 2026, bringing researchers, technology executives, policymakers and critics to Hanover, New Hampshire. The three-day gathering put human judgment at the center, with opening-day speakers addressing both how AI should develop and who should shape its place in society.

The event marks 70 years since Dartmouth hosted the summer workshop that launched artificial intelligence as a research field. Its opening day paired a keynote conversation with Thinking Machines Lab co-founder and CEO Mira Murati with an address by Yann LeCun, who received the Dartmouth AI at 70 Award.

Tools people can shape, machines that can anticipate

In conversation with Dartmouth President Sian Leah Beilock, Murati discussed the customizable AI systems her company is building. She described success in terms of keeping human flourishing, judgment and will central to increasingly capable technology. Her stated goal was to give people tools that help them define their own futures.

Murati pointed to ChatGPT’s release during her time as OpenAI’s chief technology officer as a defining experience. Technology and design together helped people discover practical uses, she said. Seeing those uses also exposed problems and misuse more clearly, giving researchers information they could use to build stronger safeguards.

LeCun focused on a different technical direction: world models, which aim to let an AI agent anticipate the consequences of its actions. He argued that this ability would help systems take safe, meaningful actions in the physical world. He pointed to domestic robots as the visible sign that such a revolution had arrived.

“Despite what you might hear from some of the more optimistic CEOs of Silicon Valley, we’re not anywhere close to human-level AI.”

Control is already a concern

New York Times Opinion columnist Ezra Klein, appearing virtually, addressed alignment: whether an AI system’s actions match human values and intentions. His concern did not depend on machines already surpassing people broadly. He argued that difficulty managing today’s systems raises a harder question about maintaining control when systems become more powerful and more numerous.

Sen. Kirsten Gillibrand called for a worldwide AI summit and more regulation in the United States, according to WCAX. She cited Congress’s failure to regulate the earlier generation of internet platforms as a mistake. At the conference, she argued for a framework that prevents foreseeable harms while allowing benefits in fields such as healthcare and education.

Beilock framed the university’s role as helping set direction rather than leaving that task to industry or government. Before her conversation with Murati, she urged participants to acknowledge that there was no established playbook and to use the gathering for difficult conversations about where AI should go.

From debate to student prototypes

The first day’s sessions extended those questions into medicine, biomedical research, energy, race and writing, and international security. Other discussions examined social change and rural economic development. The October 30 program is scheduled to include university presidents discussing institutional change and economist Bruce Sacerdote presenting research on how AI affects college students’ thinking, learning and work.

More than 140 students from Dartmouth and other institutions registered for a hackathon running through October 31. Their task is to turn questions about AI and human judgment into working prototypes. Dartmouth says community members may attend conference events in person without registering, and events are also available by livestream.