From core cognition to job readiness: Profiling AI capabilities for real-world tasks - Marko Tesic

​As AI systems are increasingly deployed in complex, human-centered environments, questions arise about what makes them ready for real-world work.

This talk introduces a cognitive profiling framework for AI systems that models their performance as a function of core cognitive capabilities meeting task demands. Using hierarchical Bayesian measurement layouts, we infer capability profiles from benchmark performance.

By aligning AI capability profiles with job demand profiles derived from human work activities, we can estimate “job readiness” scores, indicating when an AI system is worth testing or adapting for a particular task. This approach offers a principled way to track and anticipate how AI reshapes the nature of work.

Marko will be staying after the talk to discuss his work over dinner. RSVP helps us have a sense of how much food to order!