How can we live with AI?
Details
NOTE: the pub does not charge for our space on the understanding that we all purchase something from the bar. Please support them.
I'll be there from around 7pm if anyone would like a social chat before we get started.
AI is bringing huge changes to our society. Examples include: voice control of machines; automated vehicles; conversational access to the world's knowledge; medical diagnosis; and drug discovery. These are bringing substantial benefits.
But many people are concerned about future developments of the technology. How can society absorb the ever-increasing development of AI? What issues do AI developments raise that we should be concerned about? What can be done to mitigate these concerns?
We’ll start by identifying some specific concerns of interest to attendees, then think about what, if anything can be done to allay those concerns?
Concerns we might discuss include, but are not limited to:
- How can we manage changes in employment opportunities?
- How should copyright be protected?
- What steps need to be taken to ensure deployed systems do not suffer from biased training data?
- Who is responsible when AI makes mistakes?
- Will important human skills be lost, (for example, the ability to summarise information)? Does this matter?
- How should our education change? What does it mean to write an essay?
- How will we know what is “real” in a world of hallucinated facts and deepfake videos?
- Which information gets propagated? Who decides? On what basis?
- Will we get too attached to our AI “friends”, at the expense of human to human friendship?
- What forms of surveillance should be permissible – face detection, tracking…?
- Should military applications of AI be regulated? How? Eg autonomous lethal weapons, drone swarms…
- What will happen when machines get smarter than us?
If you can do some preparation, it would really help us have a more informed conversation. Here are a few suggestions:
- Interview with Geoffrey Hinton (Nobel prize winning “Godfather of AI”) on the dangers of AI. The full interview is 90mins, a shorter version is 20mins
- https://plato.stanford.edu/entries/ethics-ai/
- Slaughterbots video: A simulation video warning of what could happen in the very near future with autonomous lethal weapons.
- Character.AI bans under 18s from conversing with their AI companions
