Psychology Professor Explores the Influence of Power Systems on the World
The Troland Research Award will further Molly Crockett’s work on understanding how people learn and decide
In Molly Crockett’s lab, the researchers examine big questions about society, power, and morality. Crockett is a professor of psychology in Princeton’s Center for Human Values and director of the Future Values Initiative, which studies topics including how advances in science and technology have outpaced careful inquiry into their ethical implications.
“I’ve always been interested in people and the social world and politics and ethics,” Crockett says. “I was excited to apply scientific methods to studying these questions.”
Now, Crockett, who uses the pronoun they, will be able to devote even more time and resources to this research. They recently were awarded the Troland Research Award, which offers $75,000 “to further empirical research within the broad spectrum of experimental psychology.” The National Academy of Sciences recognized Crockett specifically for “their pioneering contributions to a mechanistic theory of moral cognition.” Computer science professor Sebastian Seung also won the award.
“I just think this is a really exciting time to think about core practices in our field,” Crockett says. “And the stakes feel really high.”
Something Crockett has been thinking about — and studying — a great deal in recent years is how generative AI is changing the field. Many scientists are excited about AI’s potential to streamline scientific processes, but Crockett and others worry about what that could mean for not only science but also how we understand and relate to human behavior.
Quantitative, computational approaches of course have value, Crockett notes.
“But my worry is that that part of our science might become the victim of its own success,” Crockett says. “If we don’t reserve more humanistic, qualitative, interpretive approaches to trouble and unsettle and challenge the quantitative approaches, we start to create really big risks for understanding over the long term.”
For example, if an AI model has results that use only quantitative means to summarize human behavior, that would be leaving out the human part of it.
Crockett’s lab explores these questions using a variety of methods, including behavioral experiments as well as computational modeling and machine learning.
“I think it’s just as important to support the interpretive, humanistic work alongside the science, because the two approaches really need each other and thrive when they’re in conversation with one another,” Crockett says. “The science and the technology are stronger when they are developed in a human-sensitive context.”
Crockett adds, “I do think Princeton really has potential to be world-class in this effort because we have so much talent in the social sciences and the humanities. Our hope is that we can develop this over the next few years.”
With the new award, Crockett has been thinking a great deal about the limitations of traditional methods, especially those that seek to quantify human behavior with mathematical models.
“The questions I’m interested in are how to understand the ways that knowledge production is changing now that there are tools that many people believe can replace knowledge,” Crockett says. “I don’t think that they can. I think those claims inspire us to get a lot clearer on what human knowledge-making actually entails.”
New tools like AI models are powerful, but when researchers try to capture something as complicated as human morality with mathematical models, inevitably things will be left out, Crockett says.
“I’ve started to think a lot about what is getting left out, and what are the consequences of leaving those parts out?” Crockett says.
“I just see so many limitations in quantitative, mathematical approaches to capturing these qualities and capabilities of humans, and so I’m worried about what paths we might go down if we forget that models are models and not reality,” Crockett adds. “How might human beings change if we take the mathematical model as the same thing as truth or reality?”
Another topic Crockett studies is social media. A decade ago, researchers studying social media and political behavior accurately predicted many of the ways the platforms are used today. For example, posts that cast someone as a clear villain generate the most engagement.
“But of course, reality is way more complicated than the stories that we read about in books or watch in the movies,” Crockett says. “I’m thinking a lot about how consuming media that portrays really complicated political events in terms of, ‘There’s a good guy, or a bad guy,’ how that simplified frame can cause us to see our own lives through that simplified frame — and what are the consequences of that?”
It can be challenging to be a scientist working on social issues, requiring researchers to navigate a fine line between sticking to the science while upholding the culture of science and engaging with a culture that often discourages political involvement. “That is unfortunate, because often it is scientists and scholars who are the most qualified to weigh in on these kinds of issues because of our expertise,” Crockett says.
All scientists, Crockett says, should value what is good about collaboration with other human beings in science or any scholarly pursuits. The goal of science is not simply producing facts about the world or creating new scientific papers but also training new scientists.
Crockett adds, “I am trying to facilitate thoughtful and critical conversations in this space.”
No responses yet