🚘 Robotaxis keep riders safe on the road, but what about their data? A USC ISI-led study finds current methods for choosing privacy-enhancing technologies fall short in real-world scenarios. Read all about it here --> bit.ly/3KsGzQh

👏 Congrats to students Zhiyi Chen, Jinyi Ye, & Beverlyn Tsai on their Best Student Paper Honorable Mention at ACM Conference on Hypertext! #HT2025 Study found 12% of election images on X were AI-generated; spread mostly by 10% of users; AI posts even drew more positive & less toxic responses.

🚀 Curious about the future of tech? Come tour USC’s Information Sciences Institute during #LATechWeek! 🗓️ Oct 14 • 10:30 AM | Marina del Rey ❄️ Quantum computer colder than space 🛰️ Small satellites + docking tech tested on the ISS 🛡️ Cybersecurity testbeds shaping the future RSVP 👉 techweek.isi.edu

Just a few days left to vote for ISI's Managing Director Terry Benzel who is running for Vice President of the IEEE computer society. Watch her candidacy interview: www.isi.edu/news/972882/...

USC ISI is hosting panels, lab tours, and a rooftop mixer during LA Tech Week 2025 🎉 Free + open to the public, Oct 13–19 in Marina del Rey. Come see the future of AI, space, quantum, and more. Details + RSVP 👉 bit.ly/46kfvKs

The fastest way for the algorithm to influence you? Go through your friends. @isi.edu researchers found that they’re often more influenceable, and that shapes your feed.

👏 USC’s ISI, w/ @UTK @UMassAmherst & @UNC, has been awarded $5M by NSF to develop PegasusAI—an AI-driven #workflow system designed to help scientists manage & execute complex research across #edge, cloud, & #HPC environments. @usc.edu @viterbischool.usc.edu #AI #Cyberinfrastructure bit.ly/45xnK5g

In July, USC ISI and @stanfordhai.bsky.social came together to spotlight the 2025 AI Index Report, with a panel that brought together leading voices from across the AI ecosystem. Read the highlights: bit.ly/45bQgdV

A study from the Uni. of Applied Sciences and Arts of Southern Switzerland and USC ISI found that Google’s Perspective API consistently rates German-language texts as significantly more toxic than identical content in English, raising concerns for content moderation. Read the story: bit.ly/46FIMB3

ISI research reveals the Susceptibility Paradox: your friends are more influenceable than you are! This award-winning work at #ICWSM2025 sheds light on why trends & misinformation catch on online. Read more: bit.ly/45aq02q #SocialMedia @usc.edu @viterbischool.usc.edu @csatusc.bsky.social

Ten ISI papers are being presented in Vienna this week at #ACL2025! From #AI teaching and safety to authorship and mime, our researchers are pushing #NLP forward. bit.ly/454O2vt @usc.edu @viterbischool.usc.edu

Thoughtful insights from Yolanda Gil, Principal Scientist & Senior Director for Strategic Initiatives in AI and Data Science at ISI, featured in the LA Times story on the Trump administration's new AI Action Plan: www.latimes.com/politics/new... @usc.edu @viterbischool.usc.edu #AI @latimes.com

New research by ISI’s Mayank Kejriwal (Daniel J. Epstein Dept. of ISE) shows male ASEAN (Association of Southeast Asian Nations) students are 3x more likely to drink than females. Read more: bit.ly/4o1kM1t #GlobalHealth #PublicHealth @usc.edu @viterbischool.usc.edu @universitypress.cambridge.org

Wrapped up the AI Index panel with rooftop views and great company! Thanks to everyone who joined us for some sunshine and refreshments after a packed afternoon discussing the most critical trends in AI. Read @stanfordhai.bsky.social's 2025 AI Index Report: hai.stanford.edu/ai-index/202...

Packed room for today’s AI Index event! Co-hosted by @stanfordhai.bsky.social and USC ISI. We had a fascinating panel featuring Ray Perrault, Yolanda Gil, Karl Jacob, Karina Montilla Edmonds, and moderator Vanessa Parli. Thanks for joining us! 🔗 More about the AI Index: hai.stanford.edu/ai-index

ISI/nsiders, ISI's podcast, is hosted by AI Division Director Adam Russell. ⁠ In this episode, Steve Crago, Associate Director at ISI, shares his experiences in how preparation and luck relate to success. ⁠ Listen here: bit.ly/44Velp6

🚨 New research from ISI, @annenberg.usc.edu & @viterbischool.usc.edu tracks real-time impact of #AI on #jobs. Some roles (like software engineer, maintenance tech) are seeing high exposure; others (HR, legal) much less. But more AI doesn’t always mean more pay. Full story 👉 bit.ly/4nxulFf @usc.edu

Join us tomorrow for a cybersecurity seminar w/ Jade Sheffley, a Ph.D. student at UMass Amherst’s SPIN lab, introducing CensorLab, a new emulation platform for studying and testing tools to fight Internet censorship. More info here --> bit.ly/3I7q6ji

Big news from #ICWSM2025! "The Susceptibility Paradox in Online Social Influence” by @luceriluc.bsky.social, @jinyiye.bsky.social, Julie Jiang & @emilioferrara.bsky.social was named a Top 5 Paper & won Best Paper Honorable Mention! 👏 Congrats to all!

New ISI study looked @ 3.5M tweets abt abortion. Liberals & conservatives used different arguments but matched each other’s hostility. One place they overlapped: talking about exceptions related to health & safety Read more: 🔗: bit.ly/44kEJHE #AbortionDebate #Polarization #AI #ICWSM

ISI/nsiders, ISI's podcast, is hosted by AI Division Director Adam Russell. ⁠ In this episode, Terry Benzel, the Associate Director of ISI, discusses discusses why fostering connection and communication is so vital to teamwork.⁠ ⁠ Listen here: bit.ly/4li1bI8

📢 You are invited: AI Index on July 17, 2025 at 4PM Location: USC Information Sciences Institute in Los Angeles PANEL AND NETWORKING EVENT Hosted by Stanford AI Index and ISI 📅 REGISTER NOW: lnkd.in/gNEfEqD9

🇩🇰 Catch our work in Copenhagen this week! From #AI & labor to #TikTok in crises, #ISI researchers are presenting 11 papers at #ICWSM25, all asking the question: how do online systems shape our lives? More here --> bit.ly/3ZLxfMi @icwsm.bsky.social @usc.edu @viterbischool.usc.edu

Joe Kemp retired last week after 42 years at ISI. He started as a computer operator and spent the last three decades supporting the Networking and Cybersecurity Division. A steady, generous presence. Congrats on your retirement, Joe! You'll be missed!

An international team of astronomers, using a neural network trained on millions of synthetic black hole simulations and data from the Event Horizon Telescope, has uncovered new insights—thanks to advanced computational support from ISI. Read the story: bit.ly/44e5AGf

ISI/nsiders, ISI's podcast, is hosted by AI Division Director Adam Russell. ⁠ In this episode, Craig Knoblock, the Keston Executive Director of ISI, discusses the differences between running a company and a research institute.⁠ ⁠ Listen here: bit.ly/43Xspg4

An ISI study led by Principal Scientist Fred Morstatter found nearly 40% of AI “facts” contain bias, a stat cited in this new piece on Hirundo, a startup tackling AI hallucinations with machine unlearning. ISI research is making waves! 🔗 Story here: bit.ly/43VLnng 🔗 Even more here: bit.ly/4e20EaW

It was a pleasure to host Visa leadership at USC ISI's Quantum Computing Center (QCC), where we house our D-Wave Advantage quantum computer. Thank you for engaging with our quantum computing research and discussing quantum’s future in finance, cryptography, and beyond.

Great first day at the FaceBase 2025 Community Forum, hosted at USC’s Information Sciences Institute! The sessions focused on integrating clinical tools and #AI into #craniofacial pain research, featuring work from USC and UCSF, and poster presentations. @viterbischool.usc.edu @usc.edu

ISI researchers used Medicare & census data to find out where doctors go after med school. Grads from top schools were less likely to work in underserved areas. Full story here! bit.ly/3FD32YK @usc.edu @viterbischool.usc.edu

ISI/nsiders, ISI's podcast, is hosted by AI Division Director Adam Russell. In this episode, Yolanda Gil, Senior Director for Strategic Initiatives in AI and Data Science at ISI, shares why there’s reason to stay optimistic about the future. Listen here: bit.ly/4dD0ir6

🧠 Can AI understand when it’s breaking the law? ISI researchers tested #LLMs on legal questions, including U.S. biological weapons #law. The results? Sometimes the models followed the rules. Sometimes they gave risky answers. Read about it here: bit.ly/4koooID

Congratulations to the ISI Class of 2025! 🎉 Yesterday, we celebrated the remarkable accomplishments of our graduating Ph.D. and M.A. students. Wishing our graduates the very best on their next adventure. From all of us at ISI: congratulations and good luck!

Join us tomorrow for an AI seminar! Ruolin Li is a Gabilan Assistant Professor at USC. In this seminar, she'll examine a series of representative transportation scenarios to explore how AVs can be leveraged to improve traffic efficiency and overall system performance. Join here: bit.ly/3GzD2Oi

ISI/nsiders, ISI's podcast, is hosted by AI Division Director Adam Russell. In this episode, Joe Barnes, a Senior Research Engineer at ISI, describes how thinking of concepts in terms of systems has extended to parts of his life outside of work. Listen here: bit.ly/4k0fq3H

ISI researchers are presenting 9 papers this week at #NAACL2025 on topics ranging from creativity, reasoning, human alignment, and more! 🔗 Read more: bit.ly/4jGOvKv @viterbischool.usc.edu @usc.edu @csatusc.bsky.social #NLP #AI

ISI’s Luca Luceri partnered with Financial Times to uncover a coordinated online campaign influencing Canadian voters. Read about it here: bit.ly/3Sb598U @luceriluc.bsky.social @usc.edu @viterbischool.usc.edu

From detecting information operations on social media to enhancing smart city development, this week ISI researchers return to the 2025 ACM Web Conference to present a number of innovative papers. ⁠ Check out the research spotlights: bit.ly/435bQzE

New research from ISI, the University of Iowa, and Claremont Graduate University sheds light on how the presence of this growing emotional divide—known as “affective polarization”—shapes individual decision-making across society. Read the story: bit.ly/3RuLj8t

Join us tomorrow for a NL seminar! Skyler Hallinan is a Ph.D. student in Computer Science at USC. He'll introduce N-Gram Coverage Attack, a membership inference attack that relies solely on text outputs from the target model, enabling attacks on completely black-box models. Join here: bit.ly/4cKP2IF

Join us for an NL seminar today! Hao Zhu is a postdoctoral researcher at Stanford University. In this seminar, he'll outline three foundational approaches developed to address challenges in training language agents. Join here: bit.ly/42iDI31

Do countries work together to manipulate public opinion across social media platforms? A new study from ISI investigates whether inter-state coordination exists using state-of-the-art detection methodologies and control datasets. Read the story: bit.ly/3EpXy30

Join us tomorrow for an AI seminar! Dr. Regli is a computer scientist who has focused on interdisciplinary and use-inspired problems. He'll summarize his personal views regarding AI he has developed during his decade away from academia in various forms of public service. Join here: bit.ly/4j3FcUP

Join us this Thursday for an NL seminar! Tolúlọpẹ́ Ògúnrẹ̀mí is a Computer Science PhD candidate at Stanford University. In this seminar, she'll discuss investigating the output of the modality adapter of spoken language models. Join here: bit.ly/4cmBQJL

Join us for an AI seminar! Nathan Lambert is a Senior Research Scientist and post-training lead at the Allen Institute for AI. He'll present historical context on language modeling and new research with RL to forecast how the industry of language models may change in the future. Join: bit.ly/3LiIMuL

Cloudwalkers in now on Amazon Prime Video: www.primevideo.com/detail/Cloud... CAST HIGHLIGHT: this week, watch our interviews about Keith Uncapher, The Visionary Founder of ISI youtu.be/0GWwxVlTSo8?... @viterbischool.usc.edu @usc.edu

ISI/nsiders, ISI's podcast, is hosted by AI Division Director Adam Russell. In this episode, Federico Spedalieri, Managing Director of the Quantum Initiative at ISI, shares why physics appealed to him from a young age. Listen to the episode here: bit.ly/42sozMn

Our next AI seminar is this Friday! Audrey Olivier is an Assistant Professor of Civil and Environmental Engineering at USC. In this seminar, she'll introduce enhanced algorithms based on ensembling with anchoring for approximate Bayesian learning of neural networks. Join here: bit.ly/43Zjdta

#Space, #Cybersecurity & #AI — SCYWAI, a 2-day workshop hosted by USC, ISI, sponsored by NSF, brought together defense leaders, engineers, investors, and researchers to tackle one urgent question: how do we secure space systems in the age of AI? 🔗 bit.ly/42sm8bN @usc.edu @viterbischool.usc.edu

Join us this Thursday for an NL seminar! Julie Kallini is a second-year Ph.D. student in CS at Stanford University. In this seminar, she'll introduce MrT5 (MergeT5), a more efficient variant of ByT5 that integrates dynamically shortens input sequence length. Join here: bit.ly/427UMYc