Computer science student at UC Berkeley develops tech to combat social media harms
BERKELEY – It's no secret that long-term social media use is not good for an individual's mental health.
Earlier this week, California joined more than 30 states in filing a federal lawsuit against Meta, alleging the parent company of Facebook and Instagram are designed to be addictive and harm children's mental health.
The lawsuit was filed by 33 state attorneys general in a San Francisco federal court.
Now, 20 year old UC Berkeley student, Sana Pandey wants to do something about the negative feelings users, young and old, get when they mindlessly scroll on their phones.
Pandey is developing AI technology that changes the way recommender systems used by Google, Facebook and other companies filter out what a user sees on their feed.
Recommender systems filter out content based on a user's reaction to certain images, videos or language. Often the subjects that are negative or make you angry get the strongest reaction.
Pandey's mission began a few years ago, when she was one of the few teens not on social media.
"It just never really called to me. And then when COVID hit, I ended up wanting to stay in touch in a way that wasn't just phone calls," she recalled to CBS News Bay Area.
Right away, she noticed a change in how she saw the world, and herself
"I realized all the information I was getting about current events was through a very specific lens and that was changing how I thought the rest of the world thought about it. And then also when I was seeing all these beautiful women all the time, I didn't realize at the time that there was editing or things that were going on behind the scenes to make them look that way," Pandey said.
Mark Nitzberg is Pandey's professor and the Executive Director at the UC Berkeley Center for Human Compatible AI.
"What her algorithm is doing is taking that, literally asking users somewhat randomly, how they feel and how they feel about their experience online last week or how they feel in general. These are carefully crafted questionnaires, and they take the output to the questions and feed them back into the algorithm. When you put that back into the system, it turns out that the outcome improves your well-being," Nitzberg said.
Pandey couldn't say which companies buys her AI algorithm because of lab contracts.
But she did say, "working very closely and collaboratively some of the biggest names out there in social media and they are all very invested in creating a better and more kind of long- term beneficial product for their users."