27 November 2017: Petros Papapanagiotou

Artificial Intelligence to support Human-Machine Collaboration:

Ideas and Directions

Hybrid human-machine systems, or the so-called “Social Machines”, are in the rise, blurring the distinctions between human and machine computation, and stretching the boundaries of what can be achieved by both. Meanwhile, people tend to focus all their attention on Data, neglecting the Processes that make scalable, trustworthy human-machine collaboration feasible. In this talk, I describe the current state of affairs regarding the modelling of collaborative processes and identify key trade-offs. I then suggest 3 ideas to leverage (mostly classical) AI techniques in this field: (1) design and development of social machines for everyone, (2) declarative terms of use and adaptive processes, and (3) formally verified smart contracts in blockchains. The ultimate goal is to crystallize these ideas in the form of research proposals.