Augmented Intelligence Workshop

This part of the workshop will focus on the ways in which people collaborate, coordinate, and influence one another to solve problems at either the individual or group level.

some organizing questions

• Coordination strategies: How can agents in a social network best coordinate to solve problems? What is known about the strategies that artificial and natural agents use to coordinate, and what are the successes and pitfalls of these strategies? How does the performance of different strategies interact with the nature of the problem being solved, the number of agents, and the benefits of exploration vs exploitation? How can an effective division of labor and specialization be achieved within a community?

• Infrastructures to support coordination: What infrastructural supports exist to allow agents to better coordinate with each other? What are the important kinds of infrastructural support: formal and informal discussion, laws and rule systems, technological aids for facilitating the sharing of ideas, online forums, competitions, and peer vs teacher instruction. What else? What are the prospects for communities developing their own infrastructures for supporting coordination?

• Network structure: How should social networks be structured to facilitate problem solving in a community of agents? What are the effects of density of connections, modularity, clique-ishness, hierarchy, and other network properties on network performance in different contexts? What are the prospects for network structure being a self-organized outcome of interacting agents?

• Spread of (mis)information: At a social scale, how does information and misinformation spread in a group? How do network homophily, filter bubbles, and motivated belief formation modify information spread? How can social networks be constructed in a way that improves the spread of accurate information?

• Mutual interactions between individual cognition and group structure? How does a group both depend on its individuals’ cognitive capabilities at the same time that it shapes these capabilities? How do individual differences in cognition produce different group-level patterns? And conversely, how do group characteristics end up producing different individual cognitive capacities and strategies?

• Collective algorithms: How do decentralized and distributed algorithms for solving problems differ from traditional approaches? What are the strengths and weaknesses of swarm algorithms? Do biological systems suggest algorithms that can be applied to problems from superficially dissimilar domains?

Recently, several technologies have been invented that have the explicit aim of helping people better solve problems and develop ideas, either by directly partnering with people or by facilitating human-human interaction. This part of the workshop will focus on technologies that aim to extend human capabilities.

some organizing questions

• Combining human and machine strengths: How can the efforts of people and machines best be combined to produce creative and effective solutions that neither would be able to generate on their own? How important is it for people and machines to understand each other to create effective teams? What will it take for people to be able to trust the solutions from AI systems?

• Computer-assisted learning: How can new technologies best be used to promote human intelligence? Should we be aiming to have computer systems provide scaffolds for constructing stand-alone human knowledge or systems that are designed to work interactively with people? What are the prospects for automated, interactive tutoring systems that adapt to their learner?

• Effective AI for humans: How must AI that is designed to be used with humans be different from stand-alone AI? For AI systems that are going to be good partners for people, what is the role of explicability, justification, computational models of human thinking, incorporating human perceptual and cognitive constraints, values, and emotions?

• Machines for connecting people with each other: What technologies do and should exist for positively and efficiently connecting people to each other? How can machines help people develop and assess new ideas, share solutions with each other, and promote productive conversations?

• Interfaces: How can high-bandwidth interfaces be constructed between humans and machines to create genuine cognitive systems at a higher-than-human scale? What are the prospects for brain-computer interfaces, simulations, and immersive experiences?

• Machines as supporting versus partnering with humans: What is the more promising model for thinking of human-machine teams: machines as supporting human reasoning and action or humans and machines co-creating new, genuinely cognitive systems? What are the observable signatures that a new cyborg systems has been created?