From August 10 to 21, the Tech Impact and Policy Center (TIP) hosted 23 scholars at the second annual Summer Institute in Computational Social Science at Stanford (SICSS-Stanford). Since 2018, institutions around the world have hosted these two-week institutes to help interdisciplinary scholars (graduate students, postdocs, and early-career faculty) learn the skills needed to conduct computational social science research. At this year’s SICSS-Stanford, we again focused on the study of online environments (as we did in 2025), but with an emphasis on AI systems as both a topic of research and a research tool.

Foundational lessons and bridging gaps

Each day mixed lectures and hands-on workshops to set participants up for success on a cornerstone of the program: group projects. Alongside our standard sessions on computing tools, ethics, and data collection, we dedicated time to AI as a research topic. Specifically, we examined topics ranging from integrating human values into AI development to large language models (LLMs) as fact checkers to AI sycophancy. We also covered AI as a research tool, with sessions on using LLMs to improve social simulations, analyze video data, and aid in generating reliable qualitative annotations. More details about these sessions can be found on our program’s schedule.

One of our aims at TIP is to bridge the gaps between research, policy, and impact. Accordingly, the 12 guest speakers who joined us over the course of the program were drawn deliberately from academia, industry, law, and policy. Participants heard from researchers who study online platforms and AI systems, an engineer who builds them at a social media company, and a legal scholar working on how they should be governed. We are grateful to these speakers for making the time to help us train the next generation.

Participants and group projects

Our 23 participants came from 19 institutions across the United States and abroad, including Northeastern, Maryland, UC Berkeley, CU Boulder, Rutgers, and the Chinese University of Hong Kong, and universities in Germany, Poland, and Austria. These participants also came from about 15 disciplines, including communication, sociology, political science, engineering, law, social psychology, and education.

For the group projects, participants formed seven teams to conduct research based on shared interests. One team began building an index of media bias indicators designed for the general public rather than academics. Another group used simulated conversations to ask whether introducing a bridge-building voice can soften the effect of polarizing ones. Other groups examined a broad range of topics, including: multilingual disinformation on Telegram, how online communities discuss and judge violations of social norms, how defendants are described in criminal trials, whether a good conversation with a stranger shapes what someone talks about next, and whether public predictions help bring about the outcomes they forecast. All groups experimented with using AI as part of their research workflow. As with our 2025 cohort, many of these teams are continuing to pursue their projects beyond our program, and we are excited to see what they produce.

Putting the “social” in computational social science

The two weeks left plenty of room for fun. To facilitate building community, we hosted an early group dinner in Palo Alto, and over the course of the institute, participants organized dinners and Bay Area excursions of their own, including a trip to the Computer History Museum a couple towns over. We closed the two weeks out with farewell drinks at Treehouse, a Stanford campus staple.

Reflecting, and looking forward

We are thrilled to have received a lot of enthusiastically positive feedback from participants since the institute ended. One participant called the two weeks “incredibly inspiring and useful, and definitely one of the best experiences I’ve had throughout my PhD.” Another wrote “I learned a lot over this time and appreciate [the organizers] being so generous with [their] time and wealth of knowledge.” Others praised the “strong sense of support and community” that we fostered, described the program as “an incredibly valuable and well-run two weeks,” and noted that the workflows we shared have “increased the productivity of [their] work.” Another simply said “C’est très bon!!!!! Absolutely phenomenal!” We concur.

We could not have asked for a better group. Our thanks go to this year’s participants, who traveled from across the country and beyond to spend two weeks with us. An institute like this one depends entirely on who shows up, and this group made it a genuine pleasure to run.

We’re also grateful to the Oak Foundation, which helped with funding our program; our student assistants, Sarah Wu, Aya Salim, and Gregory Robertson; our center director, Jeff Hancock; and our center staff, Michelle Lodwick and Michelle Pualuan, for helping everything run smoothly.

SICSS-Stanford will return next summer. Keep an eye out for details, and feel free to reach out if you’d like to get involved or contribute.