Seventy years after hosting the summer workshop that launched the field of artificial intelligence, Dartmouth last week held its AI at 70: What Must Remain Human? conference tackling tough questions about the fast-growing field.
The Oct. 29-31 conference drew leading scholars, AI executives, and students who offered a range of perspectives but also remained centered on what should be protected. AI offers the promise of dramatic improvements in productivity and capacity, they said, but the things it threatens to replace—most notably the human thought required to read a book, work on a complex math problem, or write an essay—must be preserved and nurtured, particularly within higher education.
“Anyone can get a convincing answer from AI in seconds. The hard skill now is judgment. Is that answer right and complete? What are the ethical and moral considerations? Do I understand why and can I explain and defend a viewpoint in a classroom, a boardroom, a research lab? That is what we want our students to be able to do,” President Sian Leah Beilock said.
AI literacy must go beyond efficient use of AI tools and help students understand how to think about the technology, how to push and learn without the technology, how to know whether an answer is right or wrong, and how to know when to pick up the technology and when to put it down, President Beilock said at a presidential session, How Do Universities Navigate AI’s Toughest Questions: What Does it Take to Lead?
Dartmouth Trustee Will Griffith ’93 talks with Amazon Senior Vice President Peter DeSantis ’98 about the future of human work. (Photo by Kata Sasvari)
In a conversation on More Capable AI, More Human Judgment, Peter DeSantis ’98, a senior vice president at Amazon, called himself an “AI optimist” and noted that a team of five software engineers using AI can get a project done that would have required 50 engineers in the past, and at 10 times the pace.
“What’s most exciting is seeing what can be done,” said DeSantis, who majored in computer science and economics at Dartmouth and now works on AI models, custom silicon, and quantum computing.
At the same time, he said, it’s important to “teach thinking” and to make clear that “we want to still value teaching and learning because that’s what makes us us.”
Michael Breen ’02, head of strategic initiatives at Anthropic, also spoke to the technology’s transformative potential, as well as some of the implications of such change.
“One of the most profound promises of powerful AI is it can make the capacity that’s at the heart of real, actualized human agency much more common,” Breen said in a conversation about Human Agency in a World of Powerful AI.
Michael Breen ’02, head of strategic initiatives at Anthropic, left, talks with Tuck School of Business Dean Matthew Slaughter. (Photo by Kata Sasvari)
The technology can make expertise and capabilities more widely accessible, helping people navigate systems and advocate for themselves more effectively, Breen said. But realizing its potential in practice hinges on working with researchers to address risks that come with the territory, such as a dependence on AI that can weaken human capacities, an undermining of people’s expertise, and a concentration of power among those who control the systems.
“We’ll need better answers to these problems,” Breen said.
New York Times Opinion columnist Ezra Klein said he would be “pretty skeptical” for now about bringing AI too heavily into the educational space and process.
“It’s not that people won’t use it on their own, but I think that the value is going to go back, in pretty deep ways, to say, having the attention span to read a book,” Klein said in a session Oct. 29.
Laura Carlson ’87, president of the University of Delaware, cautioned against treating writing as simply an end product by itself. “Writing helps me think. So I think there’s a very blurry line between thinking and writing,” she said.
The college years are precious and best spent in finding one’s interests and diving deep to really understand them, Mira Murati, Thayer ’12, co-founder and CEO of Thinking Machines Lab, said in a talk during the first day of the conference.
“I don’t think it’s a very good use of time to spend the next four years thinking about what jobs will now be taken by AI,” Murati said. “The point is in pursuit of knowledge and how that changes your experience and your relationship with the world, with information, with people around you.”
Murati recommended that students jump on opportunities to work on practical projects with people, particularly those with different perspectives, “because that’s the closest simulation to real life work experience.”
Quote
Anyone can get a convincing answer from AI in seconds. The hard skill now is judgment.
Similarly, in response to a question about the purpose of a university, Alastair Borthwick, Tuck ’93, executive vice president and CFO, Bank of America, said that he took away from business school the value of teamwork and collaboration, while understanding that there was room to always learn and improve.
Borthwick believes that the importance of the human touch will only increase. “I think AI, as it begins to automate and make things a little more efficient around manual work, probably puts more premium in the future around people skills, things that AI just doesn’t have,” he said.
Questions posed
Other speakers raised consequential questions about the potential impact of AI on research and human knowledge. In a talk on The History of AI and the Future of Human Knowing, Molly Crockett, a professor in Princeton University’s Department of Psychology and University Center for Human Values, took aim at the “hype”—driven by the tech industry in collaboration with the U.S. government—that is fueling investment in AI at the expense of academic science.
“Their vision is, I believe, based on a profound misunderstanding of how science works, and I think we all need to understand and be aware of what is happening,” Crockett said.
The new model centers the values of industry, productivity, and profit, and is at odds with the goal of science, which is to advance collective understanding through shared knowledge, communal values, and collaboration, Crockett said.
A discussion of race, writing, and AI Thursday Oct. 29 struck a chord with English major Emma Lefferts ’27. “There are questions about what must remain human, but there are some people who have been excluded from that category for a long time. What the effects of AI are in all communities has been definitely really valuable to think about.”
Lefferts also took away from subsequent discussions that AI is not just an abstract cloud technology; it is rooted in the labor of real communities and real natural resources. “We are talking about some predicted existential risks when there are real ones for so many communities.”
Students put the question to the test
More than 1,000 people registered to attend the conference, and 24,000 watched various livestreams. Speakers urged students to continue cultivating foundational capacities during their college years.
President Sian Leah Beilock, Anthropic’s Michael Breen ’02, and AI Conference Chair Peter Chin pose with participants in the student hackathon, including event organizers and other members of the Anthropic team, on Oct. 31, the third day of the conference. (Photo by Katie Lenhart)
To that end, the AI at 70 Student Hackathon brought together 187 students from Dartmouth and other institutions to work in teams and build, prototype, and present solutions at the intersection of AI and human judgment.
Organized by the Dartmouth AI Club with support from Anthropic, the Oct. 30–31 hackathon challenged 40 teams to explore how AI can improve people’s lives without replacing their decision-making in one of five tracks, spanning physical and mental health, education and economic opportunity, governance and collaboration, and creative expression.
The winners, Caleb Zhou ’30 and Aditya Tripathi, a first-year student at Brown University, created Wheelio, a campus navigation app for wheelchair users that finds accessible routes and entrances, and uses AI to detect and report barriers.
Other projects included an app to help Dartmouth students understand their health insurance to make informed decisions about treatment and an accessible speech-therapy and practice platform for working on lisps and building public-speaking confidence.
More research at Dartmouth planned
And the work and research continues. Professor of Economics Bruce Sacerdote ’90, who presented findings at the conference from the inaugural Dartmouth AI Survey of how college students use AI, is now surveying employers to gather information about whether entry-level jobs for college students are changing because of AI.
And the next student survey will be conducted in August 2027.
The conference was an opportunity to move beyond the abstract discussion of AI’s potential and consider some of the very real challenges associated with its development, said Jacqueline Cohen ’30, a Jack Byrne Scholar at Dartmouth who co-moderated the conversation with Ezra Klein alongside Roopika Risam, Dartmouth AI Fellow and chair of the Department of Film and Media Studies.
Cohen found Klein’s point about AI alignment and the difficulty in enforcing incredibly powerful systems to act in accordance with human intention particularly interesting.
“I am thinking more critically about that distinction between what we can build and what we can responsibly manage,” said Cohen. “I also loved getting to talk with other attendees afterward and hear their reactions, especially because people took away such different things from the same discussion.”
***
Watch videos of the livestreams from the conference, and stay tuned for other AI at 70 events.