Samuel Simko

❓Can we actually trust the quality of the existing multilingual benchmarks translated from English? Turns out many of them have some simple bugs, which hurts the evaluations - we try to fix that! Introducing Recovered in Translation 🌍 ritranslation.insait.ai 🧵below

AI is threatening our democratic society—by concentrating power, narrowing how we think, and flooding institutions faster than they can keep up. These risks emerge at the system level, and technical work alone won't fix them. 👉Check out our whitepaper with 25+ researchers: zhijing-jin.com/d/2026-ai-risk… 💡We introduce 7 threat models and ways forward. ✍️Led by @davidguzman1120 with @DaveRBanerjee, @blin_kevin, @PepijnCobben, @gcorsi_, @x_angelohuang, @ChanglingXavier, Suvajit Majumder, @psyonp, @SimkoSamuel, @strauss_irene, and @TerryJCZhang Advised by senior co-authors: @ashton1anderson, @Yoshua_Bengio, @MatthiasBethge, @RogerGrosse, Karoline Helbig, @david_lie, Richard Mallah, @radamihalcea, Susan Nesbitt, Susan Perry, @presnick, Stuart Russell, @mrinmayasachan, @bschoelkopf @audreyt and @ZhijingJin Thank you to all the institutional support from @JinesisLab @EuroSafeAI @MPI_IS @CIFAR_News @iapsAI @CARMA_411 @Cambridge_Uni @UofTCompSci @VectorInst @TorontoSRI @Mila_Quebec @LawZero_ @uni_tue @michigan_AI @UMichCSE @AUParis @UNESCO @UCBerkeley @ETH_en @ETH_AI_Center @ELLISInst_Tue @ELLISforEurope @EthicsInAI #CivicAI #AISafety #AIGovernance #Democracy #ResponsibleAI

Excited to have 3 accepted papers & 9 members of our @JinesisLab at #IASEAI2026, held at UNESCO, Paris🇫🇷! We reveal hidden authoritarian biases in #LLMs, and that fine-tuning can quietly erode model safety, exploring the risks we don't always see in AI 🔍🛡️ 🧵👇

Excited to have 3 accepted papers & 9 members of our @JinesisLab at #IASEAI2026, held at UNESCO, Paris🇫🇷! We reveal hidden authoritarian biases in #LLMs, and that fine-tuning can quietly erode model safety, exploring the risks we don't always see in AI 🔍🛡️ 🧵👇

🚨Paper Alert 🚨 Excited to release the Triplet adversarial defense at #EMNLP2025, a new approach extending circuit breaking (2024, by @andyzou_jiaming, @hendrycks et al.) which achieves < 5% ASR on embedding-level attacks! #AISafety 🔗 Link: arxiv.org/abs/2506.11938 🧵(1/6)