Why try to explain an AI's outputs when you can describe its behavior? In our #CSCW2023 paper, we introduce 𝐛𝐞𝐡𝐚𝐯𝐢𝐨𝐫 𝐝𝐞𝐬𝐜𝐫𝐢𝐩𝐭𝐢𝐨𝐧𝐬, details of model performance on subgroups of data, and find that they can significantly improve human-AI team performance.

Proud to share our #CHI2023 paper on clinician acceptance of AI treatment recommendations! We explore how in complex decision tasks, clinicians often negotiate intermediate actions using facets of an AI prediction, rather than accepting or rejecting the recommendation outright.

Next week at #CHI2023 in Hamburg: Tamir Mendel will present our field experiment with Meerkat, a social support app, testing how volunteer communities can improve mobile security. The lecture will be at Privacy on the Web session, Mon, Apr 24, 11:52 – 12:06. See you at ACM CHI Conference!

I will be at #CHI2023 in Hamburg! I will present our paper tomorrow. Eran Toch and I used Meerkat, a social support app, to compare perceptions of reliance, learning, and privacy exposure between volunteers and close connections. Text @ doi.org/10.1145/354454…

Super proud of Tamir Mendel, who presented our paper at #CHI2023, showing how our social support platform, Meerkat, is used to provide security support from people's social circles and from a community of volunteers. The paper is now out, and access is free dl.acm.org/doi/abs/10.114…

Students traveling to their first international conference

Yay oshrat ayalon!

Each time a paper gets rejected, you can't help but think "I'll never succeed as a researcher" Such negative thoughts are normal, but how can we overcome them? Our #ACL2023 📰 studies Human-LM Interaction for Cognitive Reframing of Negative Thoughts arxiv.org/abs/2305.02466 🧵

I will participate in a panel at the Digital Governance at the Age of AI Conference at Bar Ilan University, discussing ethical risks in digital devices from a data science perspective. The panel will be on Tuesday, May 16, at 14:30 at Bar Ilan University toch.tau.ac.il/participating-…

📣 Why do many existing (feature-based) explanations fail to improve human decision-making & lead to overreliance? Why do we find that example-based explanations lead to complementary performance? We have answers in our upcoming #CSCW2023 paper. 🧵👇 arxiv.org/abs/2301.07255

Comparing close connections & community volunteers in a field experiment: Eran Toch of Tel Aviv University presenting at #israHCI23 #hci

Super proud of Tamir Mendel, who I had the honor of advising in his PhD, and who schlepped all the way from his postdoc at New York University to receive his award at Tel Aviv University!

As a fellow member and colleague of UbiComp and ACM SIGCHI community, my heart goes out to Eran and other academics who are experiencing these very dark days. It’s unthinkable and I’m very sorry to hear. Please keep us posted and stay safe.

Fifteen American technology companies have already answered my call to voluntarily implement commitments that ensure AI technology is safe, secure, and trustworthy before it’s released to the public. This is vital for innovating responsibly.

WEBINAR: Assessing Israel’s Next Moves Over the weekend, Hamas militants launched an attack against Israeli towns. Israel responded with a barrage of airstrikes and is preparing a major offensive in Gaza. Our expert panel will assess the situation and explain its wider

What's going on with the AI ethics community and the Israel/Hamas war? An article at VentureBeat by Sharon Goldman dives into this painful twitter war. I was interviewed along with (((ل()(ل() 'yoav))))👾, @DanKotliar, Jules Polonetsky, Talia Ringer 🟣 🎗️, and others >> venturebeat.com/ai/ai-communit…

🚀Excited to share our curated Reading List on "Bidirectional Human-AI Alignment"! Check it out here: github.com/huashen218/bid…! Wondering what "bidirectional alignment" means? We’ve got you covered with a clear definition and a collection of representative papers in this repo.

Advice from a Doctor or AI? This #CSCW2024 paper investigates the trust in the advice source and willingness to disclose information through remote patient monitoring to receive health advice from a human doctor or an AI system. tinyurl.com/42np5s2f-cscw2…

🚨 New study published! 🚨 Featuring our very own, Dr. Devin Mann, this research explores how laypeople use & perceive Large Language Models (LLMs) vs. search engines for health queries. Key findings on trust, usefulness & accuracy. Read more: pubmed.ncbi.nlm.nih.gov/39946180/ #ChatGPT