Thinking is changing shape.
We study human cognition in AI-mediated knowledge work — what happens to attention, memory, judgment, and authorship when the systems people work with can produce fluent work on demand.
Founded 2026 Fully remote HCI · CSCW · Cognitive science
AI is no longer a tool people use. It is part of the environment they think in.
The questions that interest us are not mainly about whether these systems make people faster. They are about what thinking becomes when a capable system is always present in it.
The lab works across five directions.
-
Cognition and control in AI-assisted work
How people direct, correct, and stay responsible for work done with AI systems.
-
Attention, interruption, and resumption
How unfinished thinking survives across fragmented tools and sessions, and how context is rebuilt on return.
-
Sensemaking and reflection
How scattered material becomes understanding — noticing patterns, forming interpretations, revising beliefs.
-
Cognitive offloading and personal infrastructure
The notes, threads, and assistants people build around themselves, and how those systems start to participate in thinking.
-
Agency, authorship, and self-worth
What remains a person's own when a system contributes materially to the output, and where their sense of competence goes.
Published work
Interruptive Thinking: An Empirical Study of Operator-Side Control in AI Interaction
PPIG 2026 · UCL Interaction Centre, London · July 2026
An empirical account of how an operator interrupts, redirects, and re-enters AI-assisted thinking sessions, grounded in a dataset of classified interaction records.
Sudhanshu Garg
Founder & Principal Researcher
Independent researcher and founder. B.Tech and M.Tech from IIT Bombay. His paper on operator-side control in AI interaction was accepted at PPIG 2026 and presented at the UCL Interaction Centre in July 2026.