paper submission and info for authors | accepted papers

Panel Session

How can we avoid the risks of unreliable, manipulated, or unconstrained AI or pseudo-AI and ensure the world gets the benefits AI can deliver?

Chair Andrew Lea, University of Brighton

Panel Members

Dr. Peter Garraghan (Lancaster University)				Professor Frans Coenen (University of Liverpool)
Dr. Charmaine Sample (Idaho National Laboratory)				Clive Spenser (Logic Programming Associates)

This year’s panel session will discuss “How can we avoid the risks of unreliable, manipulated, or unconstrained AI or pseudo-AI and ensure the world gets the benefits AI can deliver?”

As ever, even in the pandemic, AI remains topical, partly because people hope that it may be part of the solution, and partly because data analytics is so often presented as “artificial intelligence”. Even formulas have now been promoted to being algorithms which, if one believes some sections of the media, have only just been invented! We see therefore that there is genuine applied AI, and pseudo-AI.

This ubiquity of AI, pseudo AI, and misused AI (if there is such a thing) brings multiple issues, which the panel will discuss this year. These issues may include:

“Bias in AI and machine learning”. Arguably, and arguing is what a panel session is for, any “bias” in AI simply relates to predispositions in the data being learnt from; the “bias” originates in the behaviours which originally gave rise to that data. Again, arguably, what is portrayed as “bias” may be simply that which is socially undesirable; certainly to be addressed, but how?
Fake news is topical, and ranges from actual fake news created for nefarious, financial, or political gain; through to perfectly genuine news which is being labelled as fake by those who oppose its implications. Can AI be used to detect this fake news? How do we prevent “AI” being used to give credence to the discrediting of genuine news, or to generating plausible fake news in the first place?
If AI could help in fighting the pandemic, it would of course be very welcome. How can AI assist in international emergencies such as Covid19, or the slower burn of climate change and global warming? Is AI, or the misuse of AI, in any way responsible for the mismanagement of the pandemic, if indeed it has been mismanaged?
If AI is used for decision making, how can we protect society from unreliable, unconstrained, or even manipulated AI?

In short, how do we ensure the world gets the benefits of AI which, as with any technology, can be used or misused for good or ill.

paper submission and info for authors | accepted papers

sgai-conference@bcs.org.uk