Mikita Balesni works on AI safety and alignment, with a focus on evaluating and mitigating risks from highly capable language-model agents. His research has covered scheming, situational awareness, strategic deception, chain-of-thought monitorability, and alignment training.
Mikita Balesni works on AI safety and alignment, with a focus on evaluating and mitigating risks from highly capable language-model agents. His research has covered scheming, situational awareness, strategic deception, chain-of-thought monitorability, and alignment training.
Career
Mikita Balesni holds a Master of Science in Computer Science from University of Tartu Institute of Computer Science.
Research Scientist and Founding Member at Apollo Research until 2025Worked on AI safety cases, evaluations of frontier models for scheming and situational awareness, and chain-of-thought monitorability. Co-authored research on in-context scheming, strategic deception, and evaluations-based safety cases.Cosigns Received
Oleg MürkAI researcher and former OpenAI technical staff member focused on formal verification, AI control, and competitive programming.
OpenAI has fired three safety researchers, alleging they mishandled confidential information. After the recent multi-agent security incidents, will the technical leadership overseeing these systems face the same accountability? I was pushed out of OpenAI in August after ~5 years as a researcher. No foul play in my case, but working on formal math & verification in Lean at OpenAI was career suicide. During my time at OpenAI, I became increasingly concerned that pressure to ship and beat competitors was taking priority over AI safety. OpenAI may have had legitimate procedural concerns. But after the major multi-agent SNAFU, firing three of your safety researchers feels just stupid. I doubt there is an evil master plan here. Just chaos, misaligned incentives, and intense pressure to ship. @sama OpenAI should consider rehiring these researchers!
Daniel KokotajloExecutive Director, AI Futures Project
Thank you for speaking up about this. I don't think you did anything to justify being fired and think that OpenAI should give proof rather than just insinuations. Reminds me of the equity situation two years ago and of leopolds firing before that.
Neel NandaMechanistic interpretability lead at Google DeepMind
This seems extremely sketchy behavior from OpenAI I don't want to pass judgement before hearing OpenAI's side, but I doubt these three are lying about not hearing good reasons And why would 3 associated safety researchers be fired *simultaneously* for 2-3 unrelated events?
Felix BinderAI Alignment Researcher and Cognitive Scientist at Meta Superintelligence Labs
I think highly of Tomek, Mikita and Jasmine. OpenAI firing them has made me more skeptical of OpenAI as an organization and OpenAI’s ability to ensure that their models are safe
Leo McKee-ReidFounder of Neolithic and AI safety researcher
Mikita was my research fellowship supervisor 2 yrs ago. He an incredibly cracked researcher, cares deeply about humanity, and has high integrity. I trust his concerns and think others should too.