Eran Toch
@eranto.bsky.social
360 followers 86 following 140 posts
I invent new technologies and user experiences for privacy & human-AI interaction. Prof of Engineering at TAU.
Posts Media Videos Starter Packs
Pinned
eranto.bsky.social
Exciting news: our paper “Differential Privacy Configurations in the Real World: A Comparative Analysis” was just accepted to IEEE Transactions on Knowledge and Data Engineering. Another great paper of Michael Khavkin’s Ph.D.! Here is a short thread about the study >>
Paper screenshot
eranto.bsky.social
I would adjust my settings asap on Venmo if I don't want to get phished by the whole world.
Reposted by Eran Toch
eranto.bsky.social
I was very happy to give a Cybersecurity Colloquium seminar at the The Data Science Institute at Columbia University, presenting with this stunning view. I was talking about: When Differential Privacy Meets Human Decision-Making.
eranto.bsky.social
I don't do reviews anymore 😭
eranto.bsky.social
What's the meaning of life after the #chi2026 deadline had passed?
eranto.bsky.social
Sounds like a great initiative!
eranto.bsky.social
By the way, I only teach this course once every three years — and each time, the technologies have completely changed 😅 www.ims.tau.ac.il/tal/kr/searc...
אוניברסיטת תל אביב-תוצאות חיפוש קורסים
www.ims.tau.ac.il
eranto.bsky.social
In the course, we’ll tackle challenges such as uncertainty, sharing control, and the difficulty of building a mental model for systems based on language models and machine learning. Practically, we’ll design interactions for such systems and build chatbots powered by language models >>
eranto.bsky.social
Everyone is working on AI, but how can we design AI-based systems that are actually friendly and usable? Bidding closes soon at TAU. so it’s a good time to remind that I'll be teaching the graduate course “Human–AI Interaction” (2 hours, open to the entire campus). Here is a short explanation >>
eranto.bsky.social
זה היה באוטובוס?
eranto.bsky.social
כן, אני יודע! סתם מחשבה שעלתה לי כשראיתי את זה.
eranto.bsky.social
זה מעניין. אני רואה איך האסתטיקה של גרטה ת׳ונברג יכולה להיות ממש משחררת ללא מעט אנשים.
eranto.bsky.social
It will be my second year co-chairing the Privacy and Security subcommittee at #CHI2026, with the awesome Florian Schaub and Emilee Rader. Abstract submission is today, and we are very excited to review the list of papers that you'll send our way...
handle.invalid
Reminder🔉 Abstract/metadata deadline is today (Sep 4 AoE)! No new submissions and author changes after the deadline. Make sure metadata is finalized before time runs out!
eranto.bsky.social
The preprint of our paper, “Differential Privacy Configurations in the Real World: A Comparative Analysis,” by (brilliant) student Michael Khavkin and myself, has been published on the IEEE Transactions on Knowledge and Data Engineering website. Read it here: www.computer.org/csdl/journal...
IEEE screenshot.
eranto.bsky.social
On Wednesday, September 3, the senior faculty at Tel Aviv University will hold a warning strike. We will head to Jerusalem to protest against the dismantling of the rule of law, the continuation of the war, and the indifference toward the hostages.
eranto.bsky.social
Our main point is that #DifferentialPrivacy is pretty great because it allows informed discussions of privacy guarantees, but configuring them is not just a technical choice, it’s a major policy decision with direct implications for user trust, fairness, and transparency.
eranto.bsky.social
We propose standardized ways to frame and report on DP configuration processes and suggest a framework for an "Epsilon Registry", a public database of DP implementations to guide applications, inform regulators, and improve transparency to users. >>
Framework for epsilon repository
eranto.bsky.social
Often, these choices aren’t clearly explained. 59% of academic papers and most of commercial deployments provided NO justification for their ε choices. This lack of transparency makes it hard to reproduce results or understand privacy guarantees. >>
Explanations for epsilon
eranto.bsky.social
All sectors show increasing ε values over time, suggesting organizations are relaxing privacy constraints to enable more fine grained data analysis. This trend raises serious questions about the long-term DPs privacy protection and whether it protect data subjects at all. >>
eranto.bsky.social
Our findings show that privacy guarantees (e.g., epsilon ε values) provided in theoretical academic research are rarely matched in commercial and governmental deployments, which adopt way looser guarantees >>
epsilon over time
eranto.bsky.social
But how is it actually deployed in the real world? And most importantly, how strong are the privacy guarantees provided to actual users? We conducted the first systematic review of real-world DP deployments (n=140), comparing commercial, governmental, and academic uses >>
Systematic review process
eranto.bsky.social
Differential Privacy is the gold standard for data anonymization as it provides mathematically provable privacy guarantees for individuals. It’s already used in products from Apple, Microsoft, and even the U.S. Census. Also the basis of privacy-preserving machine learning >>
Apple differential privacy
eranto.bsky.social
Exciting news: our paper “Differential Privacy Configurations in the Real World: A Comparative Analysis” was just accepted to IEEE Transactions on Knowledge and Data Engineering. Another great paper of Michael Khavkin’s Ph.D.! Here is a short thread about the study >>
Paper screenshot
eranto.bsky.social
Hundreds of thousands gathered yesterday in Tel Aviv calling the government to return the hostages and end the war. By the way, the demonstration happened about two hours after a Houthi ballistic missile had cleared people off the streets.
Demonstration in Tel Aviv View of demonstration in Tel Aviv