Alumni profile: Saffron Huang, A.B. '20

A trip to Peru taught Saffron Huang quite a bit about how technology, politics and economics intersect. Huang, A.B. '20, was studying applied mathematics at the Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS) with a secondary in government, and in 2019 went on a winter break trip to study the effects of illegal mining on the Amazon River.

“I realized that a lot of this is a social problem, that it’s politics and economics, and there’s not a lot of stuff you can do on the technical front if you don’t get the political will behind your efforts,” Huang said. “That made me really aware that solutions need to be interdisciplinary.”

The desire for interdisciplinary solutions to societal challenges led her to co-found the Collective Intelligence Project in September 2022. The project is focused on promoting tech and especially artificial intelligence (AI) policy, whether at specific companies or a local government.

“I’m really interested in trying to implement things on the ground in a technically and politically informed way,” Huang said. “I want to see how we get pilot programs into practice in terms of better governance for AI and using AI to improve our collective decision-making, consensus-building and governance methods.”

Huang grew up in New Zealand, which at times felt “very far away from the rest of the world.” By high school she knew she wanted to go to college in another country, eventually leading her to Harvard. She joined the Berkman Klein Center for Internet & Society while here, which is where her interest in AI began.

“Harvard is where I realized how technology is shaped, which had a big impact on me,” she said. “All my research at the Berkman Klein Center pointed to this common theme of thinking about society and artificial intelligence, how we can learn about human society by simulating it with AI.”

While at Harvard, she studied AI in other ways as well, including its impact on international relations through a summer internship with the British Foreign and Commonwealth Office. Huang later interned at DeepMind, Google’s AI research lab, and after graduation became a full-time research engineer there.

“It made me interested in how we can teach AI to be more cooperative, and understand how humans work,” she said. “DeepMind is very academic in a lot of ways, and people have lots of ideas that could not be supported by a short-term project-driven company. But DeepMind has the resources to explore things that might not materialize for 5 to 10 years. It was kind of like academia, but with product managers.”

At the Collective Intelligence Project, Huang researches how AI governance and policy can balance progress, participation and safety, referred to as the “transformative technology trilemma” on the organization’s website. Her organization’s premise is that policy doesn’t have to sacrifice any of these three elements, and if implemented correctly can encourage large numbers of people to push the technology forward without putting specific demographics at risk. Huang points to the example of Polis, a Taiwanese social media app that encourages political consensus and reasonable policy debate.

“I’m talking about the decision-making processes that people institute in order to agree on what we should do together, and what form we want things to take in our society,” she said. “We’re going to manifest this by starting with small-scale experiments on the level of specific company policies. We want to run some trials with specific platforms, and we’re also considering working on long-term proposals at the city level.”

Huang has always thought multidisciplinary approaches work best when it comes to big challenges and she applied that philsophy to her coursework at Harvard. She took government and applied math classes, of course, but also film studies classes and courses on ethical issues in data science. Three years after graduating, she’s still using that multidisciplinary approach to help shape AI’s ever-growing role in our society.

“I’m very shaped by Harvard in terms of the way I think and analyze situations,” she said. “Feeling like liberal arts matter to technology is not something a lot of people believe, but I very much do.”