This is a joint event with the Oxford Martin School and the Oxford Martin Programme on the Impacts of Future Technology
Dr Joanna Bryson, Reader, Department of Computer Science, University of Bath, will ask: Is artificial intelligence an existential threat to humanity? If so when?
In this talk she will argue that what Dewey, Goertzel and others have described as an intelligence explosion is an accurate description of the impact the socio-technical system of humanity and its culture has had on this planet since the dawn of writing – arguably the first form of AI. This pattern is accelerating with new communication and computation technology creating an ever-more powerful and dynamic system. Nevertheless, she argues that understanding this system may be within the capacity (and the remit) of theoretical biology.
She will then move to a discussion of our capacity to control the system, via altering the moral agents that compose it – that is, by providing useful abstractions to the general population and/or policy makers. The EPSRC Principles of Robotics (possibly the only national-level general-purpose AI ethics policy statement, although presently completely non-binding) calls this process "transparency" and mandates it. Expected points of discussion:
- whether referencing "the" singularity as being a future rather than a past event benefits transparency,
- whether there are other things we can do to make AI more transparent, and
- whether there is a way to test or model the impact of measures to increase transparency, both for their effectiveness in education and their impact on containing or even exploiting the present explosion.
This lecture will be followed by a drinks reception, all welcome.
This lecture will also be shown via live webcast and can be viewed here:
About the Speaker
Dr Joanna J. Bryson is a Reader in the Department of Computer Science at the University of Bath. She is an Associate Editor for Adaptive Behavior and on the Editorial Boards of the International Journal of Synthetic Emotions (IJSE), the Journal of Mind Theory, Connection Science and AI & Society.
Joanna’s principle academic passion is understanding human behaviour, human culture, and natural intelligence more broadly. Her main methodology for doing this is designing intelligent systems to model and test scientific theories. With her colleagues in Artificial models of natural intelligence, she publishes in many of the disciplines cognate to cognitive science, including anthropology, animal behaviour, cognitive robotics, computer game character AI, philosophy and ethics.
Joanna earned her B.A. in Behavioural Science at the University of Chicago in 1986, her M.Sc. in Artificial Intelligence at the University of Edinburgh in 1992, her M.Phil. in Psychology at the University of Edinburgh in 2000 and her Ph.D. in Computer Science at MIT in 2001.