Linde Center for Science, Society, & Policy at Caltech | Research & debate on topics at intersection of science, society, & policy | R.M. Alvarez & F. Eberhardt
Can AI help us breathe cleaner air? 🌍
New paper from @CaltechLCSSP's Ramit Debnath,
@Cambridge_Uni & coauthors found that a small data choice changed accuracy 50–75%. For climate AI, problem design matters as much as model design.
📖 Open Access: iopscience.iop.org/article/1…
🎉Congratulations to @CaltechLCSSP's, @DannyCEbanks & Jonathan Katz, with coauthor Gary King for turning that idea into remarkable impact:
✅ 26.9M Google impressions, 1.1M+ visits from ~690K users, Recognition from Drudge Report and Jimmy Kimmel!
dannyebanks.com/map/
Congratulations to @CaltechLCSSP's @DrLucyJones, @Caltech on launching the Center for Disaster Resilience, uniting scientists, engineers, & community leaders to advance preparedness, response, & recovery from our region’s greatest hazards. Looking forward to the impact ahead! 🔥
💡AI can accelerate development, but without inclusion it may deepen existing inequalities. @CaltechLCSSP's Ramit Debnath & @Cambridge_Uni's Abdullah Safir argue that AI’s future depends on whose voices, knowledge, and priorities are represented.
🔗universityworldnews.com/post…
New paper: "Representational Control over Self-Report & Behavior Coherence in LLM Risk-Taking"
🎯 𝗗𝗼 𝗟𝗟𝗠𝘀 𝗵𝗮𝘃𝗲 𝗮 𝗿𝗶𝘀𝗸-𝘁𝗮𝗸𝗶𝗻𝗴 𝗱𝗶𝘀𝗽𝗼𝘀𝗶𝘁𝗶𝗼𝗻 𝘁𝗵𝗮𝘁 𝗰𝗼𝗵𝗲𝗿𝗲𝗻𝘁𝗹𝘆 𝗴𝘂𝗶𝗱𝗲𝘀 𝗯𝗼𝘁𝗵 𝘁𝗵𝗲𝗶𝗿 𝘀𝗲𝗹𝗳-𝗿𝗲𝗽𝗼𝗿𝘁𝘀 𝗮𝗻𝗱 𝘁𝗵𝗲𝗶𝗿 𝗿𝗶𝘀𝗸𝘆 𝗱𝗲𝗰𝗶𝘀𝗶𝗼𝗻𝘀?
𝗧𝗟;𝗗𝗥
𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀: LLMs often say one thing about themselves and do another. In risk-taking the gap is costly: a clinical assistant or a trading agent can call itself cautious and still make risky decisions.
𝗪𝗵𝗮𝘁 𝘄𝗲 𝘀𝗵𝗼𝘄: No shared disposition: inside the model, stated risk-taking and risky choices are represented separately. Task-specific instructions seem to move both, but through misleading shortcuts like producing larger numbers.
𝗧𝗵𝗲 𝗳𝗶𝘅: Steering internal representations of stated and enacted risk together makes what a model says match what it does, without retraining.
𝗧𝗵𝗲 𝗱𝗮𝗻𝗴𝗲𝗿: The same mechanism can make a model say the opposite of what it does (report low risk while taking more, or the reverse), in every model we tested.
📄 Paper: arxiv.org/abs/2610.04125
🌐 Website: psychology-of-ai.github.io/r…
w/ @p_song1@pengrui_han@MyrlGMarmarelis, Ramit Debnath, Dean Mobbs, @rmichaelalvarez, @CaltechLCSSP and @Cambridge_Uni's Centre for Human-Inspired AI (CHIA)
Technical details below 👇
#AgenticAI#AISafety#AIAlignment#Interpretability#ActivationSteering
🗳️Today is Election Day!
Democracy isn't something we watch, it's something we participate in. Every eligible voter deserves the chance to cast a ballot in elections that are:
⚖️ Fair ✅ Secure 👥 Accessible
Whether voting in person or by mail, engagment matters.
LAST CHANCE! Sign up to attend a free event with @JessicaHuseman, @aztammyp, @rmichaelalvarez, and Bob Page, registrar of voters for Orange County, California.
link below ⬇️
✨A highlight of this week was @CaltechLCSSP Research Affiliate Ramit Debnath, @Cambridge_Uni & CHIA delivering the keynote at the #ClimateNLP workshop at #EMNLP2026 in Budapest.🌍
The presentation drew on a long-running collaboration between CHIA & @CaltechLCSSP.🤝📚
New Research Alert! @DannyCEbanks and I have been studying the dynamics of US elections for some time, here's our current perspective of persuadable voters as we head into the final two weeks of the US midterm elections!
frontiersin.org/journals/pol…
Interested in learning more about why we see "blue shifts" and "red shifts" in post-election voting returns? Give our simulation a spin, it's based on published research with @yimenglipolisci and synthetic data.
rmichaelalvarez.github.io/bl…
🎉Congratulations to @CaltechLCSSP Board Member @HannahDruck & Prakash Mishra, @Caltech for launching the PERMIT Lab with a $7M gift from the WHH FOUNDATION to advance research at the intersection of environmental economics, AI, & sustainability. 🌎🧠🌱
Learn more below:
A new Caltech lab will soon inform environmental policy. The lab is being established through a $7 million infusion of funding from the WHH Foundation, a private family philanthropic organization.
Read more: caltech.edu/about/news/7-mil…
There is still time to register for the Caltech Election Night Watch Party!
📅 November 3, 2026
🕖 7:00 PM - 10:00 PM
📍 Avery Courtyard, Caltech
Join fellow @Caltech community members and follow federal, state, & local election results together.
🔗forms.gle/fh4RV9QU2zUyxKK26
🎉 New publication in Environmental Research Letters!
Analyzing 232 Chinese cities with explainable ML, @CaltechLCSSP's Ramit Debnath, @CHIA_Cambridge & coauthors identify pathways for cities to go from high- to low-carbon growth in 10 years.🏭 🌱
🔗 iopscience.iop.org/article/1…