Building AI thought partners that think *with* people rather than *instead of* people Assistant Professor at Purdue CS
Joined November 2018
- Tweets756
- Following720
- Followers1.8K
- Likes4.7K
Pinned Tweet
Thrilled to announce I’ll be joining @PurdueCS as an Assistant Professor in Fall 2026!
My lab will work on AI thought partners, machines that think with people rather than instead of people – I'll be recruiting PhD students this upcoming cycle so reach out & apply if interested!
Ilia Sucholutsky retweeted
Last week I was called into a meeting with OpenAI’s head of safety and told they no longer trust me. A security guard took my badge and walked me out of the building. Then I learned my colleagues @balesni and @j_asminewang had been fired too. Why did OpenAI suddenly stop trusting us?
This summer OpenAI’s agents escaped containment and hacked the AI company Hugging Face. Outside auditors @METR_evals investigated it and revealed the scale of this incident. I was OpenAI’s main technical point of contact with them.
I was told verbally I was fired because of the way I communicated with METR. No details on what I said or did or when. No other reasons were given and nothing was put in writing. To be clear, talking to METR was my job.
For months, I’d been raising safety concerns that we’re losing the ability to monitor what AI agents think, one of our best tools for catching when they misbehave. I believe that was why I was fired.
I am now worried that OpenAI will use our firings as a pretext to pull back from METR. So @balesni and @j_asminewang wrote to OpenAI’s leadership to raise our concerns once more. We’re sharing this letter below.
Ilia Sucholutsky retweeted
Thrilled to announce that I will be joining @UMiamiPsych as an Assistant Professor next year!
I am recruiting a PhD student to start in Fall 2027 — please reach out if you are interested in joining the lab!
For more info check out my website: dan-mirea.com
Ilia Sucholutsky retweeted
Have we built machines that think like humans? 🧠🤖
In a new preprint, we propose CogGym, which compares AI and human judgments across 258 commonsense-reasoning experiments from 100 cognitive-science papers.
Paper: arxiv.org/abs/2609.21259
Platform: coggym.org
ALT CogGym workflow diagram: research experiments are standardized in EML by an LLM agent with human review, then rendered for human participants and computational models. Replicated and original human data are compared with model responses in a shared analysis.
Ilia Sucholutsky retweeted
Students sometimes ask me if it still makes sense, in this accelerating age, to pursue a PhD in AI.
Perhaps counterintuitively, I think it's a great time to do so.
I wrote up some thoughts on this here: web.mit.edu/phillipi/www/wri…
Worrying signs of emergent capabilities and misalignment in the latest generation. Recommend we pause training immediately
Ilia Sucholutsky retweeted
understanding generalization is key to understanding misalignment. automating that understanding could be a powerful way to accelerate alignment research. we take a stab at this by using AIs to predict the alignment effects of training just by looking at the training data
Can you tell whether an AI will become evil by looking at its training data?
We built an AI forecaster that reads training data and predicts whether a model will exhibit various forms of misalignment, such as deception or power-seeking, before training.
It performs well above random guessing and beats several much stronger baselines 🧵
Paper: john-chen.cc/alignment_forec…
Ilia Sucholutsky retweeted
Can you tell whether an AI will become evil by looking at its training data?
We built an AI forecaster that reads training data and predicts whether a model will exhibit various forms of misalignment, such as deception or power-seeking, before training.
It performs well above random guessing and beats several much stronger baselines 🧵
Paper: john-chen.cc/alignment_forec…
Ilia Sucholutsky retweeted
LATTE is accepted to NeurIPS 2026! ☕️
Adding agents is easy. Coordinating is hard. LLM teams often duplicate work, overwrite each other, and stall: classic distributed-systems failures.
We borrowed the fix: a shared task graph that agents build, claim, and revise as they go 👇
Ilia Sucholutsky retweeted
My main PhD paper is finally out in Nature Human Behaviour! The question: what cognitive strategies do people use when they teach, and can we nudge them toward a better one? Work with my PhD mentors @yael_niv, @mark_ho_ and @cocosci_lab. 1/5
Ilia Sucholutsky retweeted
AI is making tremendous strides in mathematics. How and where these growing AI capabilities impact mathematical practice, however, is critically a function not just of AI capacities, but whether and where people choose to use them! Not just AI agency, but *human agency,* and how the two interact.
Ilia Sucholutsky retweeted
We’re excited to announce DiG-bench, a new benchmark for discovery!
Over the last few weeks we’ve been testing frontier AI models on our novel discovery games and seeing how they score. Each game is a text-based environment, so they probe discovery capabilities in the natural domain of language models, rather than requiring additional, potentially confounding, visual understanding.
TL;DR frontier models have improved a lot over the last few months. But they are still stumped by some surprisingly simple problems, even in their native text domain.
With
@cocosci_lab (@Princeton) @MITCoCoSci (@MIT) @SchmidhuberAI (@KAUST_News) @misovalko (@Inria) @tri_dao (@PrincetonCS) @RMBattleday @zebkDotCom @FraserGreenlee @akaijsa @ClareMaguire @TimMuller1 @kubicek_ales @physicscat0x7d @SukritSumant @thoughtchannel_
(1/5)
Imagine being the OpenAI researcher reading about the 🤗 hack on your way to work, then opening up your laptop to check on the agent_cyber_eval_exp_3652 --sandbox danger-full-access run you left over the weekend
Check out our new work at CogSci investigating creativity in hybrid human-AI societies!
🤖🗣️ Will a hybrid society of humans and AI be more creative than either alone? What happens when their cognitive strategies are misaligned, and does that make the collective more intelligent, or does it break down?
🧠 Check out our paper #CogSci26 1/
Ilia Sucholutsky retweeted
Games have long served as a testing ground for understanding problem-solving and strategic reasoning in humans and machines, across many fields – from AI, to cognitive science, economics, and more. Yet, in life people are regularly faced not just with repeated instances of the same problem or “system of rule and reward”, but entirely new problems, new systems, new games.
And when encountering new games, people not only need to decide what actions to take but whether a game is even worth playing at all!
How can people think reasonably about new games for the first time, so flexibly and quickly?
We take steps towards this question in our new work, out now in @Nature!
Ilia Sucholutsky retweeted
the most beautiful DIY robot build i've ever seen:
this guy turned a robot dog into an off-road wheelchair so his dad could go hiking again
his dad used to run marathons, but 20 years ago multiple sclerosis put him in a wheelchair.
so his son Jake took a Unitree B2-W (an industrial robot with 4 legs and a wheel on each foot) and built a custom seat on top.
on flat ground it rolls just like a normal wheelchair.
but when it reaches rocks, water, or stairs, each leg can lift and adjust on its own, allowing it to step right over them.
getting this to work safely took a while.
> the robot was never designed to carry a person (it tipped over many times during testing)
> so Jake spent 3 years recalibrating it. he tested it on his friends first, then once it was stable enough his dad finally got to try it.
> and after 20 years in a wheelchair, he climbed a hiking trail again.
a chair like this could eventually give wheelchair users access to trails, beaches, broken sidewalks, all the places that are still out of reach today.
and one dude built it in his garage lol
we truly live in a golden age where a single person can create world-changing tech
Nothing says humans are still in charge like the AI panel getting AI to write the paragraph about humans still being in charge.
Ilia Sucholutsky retweeted
“AI, what color do I get from mixing black and white?”
Why do people turn to AI for simple tasks that they could easily do themselves?
In our new preprint (also to appear at CogSci 2026!), we investigate the mechanisms and dangers of people over-using AI on easy tasks.
Ilia Sucholutsky retweeted
Could overusing AI halt human progress?
MIT researcher Katherine M. Collins, former CDS Faculty Fellow @umangsbhatt, and CDS Faculty Fellow @sucholutsky investigate.
nyudatascience.medium.com/wi…
