SAB member professor Carolyn Rosé: “As we clarify what the frontier of capability is both for humans and for AI, it will be more clear where the sweet spot is for synergy”

Carolyn Penstein Rosé is professor of Language Technologies and Human-Computer Interaction, Carnegie Mellon University, and Language Technologies Institute and Human-Computer Interaction Institute, US. Rosé is member of international Scientific Advisory Board (SAB) of Hybrid Intelligence programme at the University of Oulu. Professor Rosé shared her thoughts on the most interesting aspects about hybrid intelligence while she visited Oulu.
Portrait of Professor Carolyn Rosé outside in springtime

Professor Carolyn Rosé´s definition of hybrid Intelligence reflects well her field of research too, and she gives a very tangible explanation for the concept. To her, Hybrid intelligence is an integration between what humans do well and what AI models do well.

“As researchers in this area, we seek to discover new ways of achieving more through a human-AI partnership than can be accomplished either by humans or by AI alone. Key challenges for current machine learned models are abstraction and problem decomposition, which humans are good at.”

The synergy between human intelligence and AI lies in the nuances

Professor Rosé brings up some of the most interesting questions in the field right now.

"In order to achieve the greatest synergy between human intelligence and AI, we must probe deeply into the nuances of what makes some tasks challenging. “

Typical existing benchmarks for AI performance obscure the assessment of capabilities because the distribution of phenomena in naturally occurring data is such that a relatively small amount of the natural variation occurs the bulk of the time, and each of the numerous other cases occur rarely. Thus, machine learned model performance masquerades at succeeding with abstraction and decomposition through memorization of specific problem decompositions.

“As we clarify what the frontier of capability is both for humans and for AI, it will be more clear where the sweet spot is for synergy", she continues.

Professor Rosé sees opportunities for Hybrid Intelligence programme to advance research on education that meets the needs of diverse learners in the future

Looking ahead, Professor Rosé shared her expectations for the Hybrid Intelligence programme over the next five years. She sees that work at the confluence of AI and the study of human learning will reveal new opportunities for impact.

“As the field of education seeks to meet the needs of diverse learners, the hope is that more powerful modelling approaches will enable nuanced treatment of individual differences, including learning disabilities, and conditions such as ADHD”.

Reflecting on her visit to Oulu, Rosé expresses her admiration for the team and their goals. "I am impressed with the well-rounded team and ambitious goals. I look forward to seeing what the team produces.”

Last updated: 10.1.2025