Since 2015, I have contributed to an international research effort to investigate the role of physical and digital learning spaces in supporting knowledge communities. Within this research, one area of interest is the design and use of “smart spaces”—i.e. physical spaces that have a degree of intelligence embedded into their architecture, such that agency within the room can be distributed among the humans, technology, and tangible/physical elements within the space.
The use of tangible and ubiquitous computing within smart spaces allows the technology elements to “disappear” into the walls, furniture, and objects in the room. In the field of computer-supported collaborative learning, this is transformative because it allows human participants to move around the space freely and to converse with other individuals without their personal devices/screens consuming their gaze. At the same time, this also provides an opportunity to build locational awareness into physical spaces, such that the room ‘knows’ when and where you are present (e.g. using locational beacons to provide personalized content to specific groups/individuals based on their roles or their current activities).
As well, within the context of smart spaces I have a deep interest in exploring the relationship of agency between humans and technology, including the application of “delegation agents” within such spaces. In agent-based systems, a delegation agent allows for the adaptive allocation of decision-making responsibility between humans and software agents, such that responsibility at any particular moment will go to the decision-maker most able to perform the task (e.g. based on an optimization of task arrival rates, waiting time, task priority, ability, and service time required to execute a task). In some situations, a human may be better equipped to serve as the decision-maker (e.g. for situations involving value judgments, ethical dilemmas, and unscripted/”black swan” events that are difficult to predict or log computationally). In other situations, a software agent may better serve as the decision-maker (e.g. for decisions based on searches of large databases or the execution of complex algorithms).
In a world where AI is experiencing rapid advancement, understanding and negotiating our relationship of agency with intelligent software agents is critical, as it will have tremendous implications on our legal system, insurance and regulating bodies, ethics protocols, and overall way of life in the future.
Alisa Acosta 