Lecture 7: Ethics of AI

Question 1

Why should one care about the ‘rise of machines’? (4)

Answer 1

Computers now have unprecedented computing abilities.
More data ( same amount of data from civilisation to 2003 made every 2 days ) makes things previously not computable into computable units.
Increased autonomy => AI agents (AlphaGo)
Takes a larger role in human life

Question 2

There are obvious benefits of AI, but what are some of the bad that could arise from it?

Answer 2

There are ethical risks which emerge from AI which may be:
Intentional: Autonomous weapons or deepfakes
Unintentional: Biased algorithms (e.g face recognition for crime)

Question 3

In what tasks has it been said that AI has overtook humans or are at least as good as them? Give examples

Answer 3

Risk management
Supply chain management (e.g Amazon supply chain logistics)
Medical diagnosis
Perfect information games
Imperfect information games (e.g No-limit Texas hold’em)
Writing poetry (???? ‘at least as appealing’)

Question 3

What does aversion to AI or appreciation of AI often depend on?

Answer 3

The task, people tend to trust algorithms when it comes to consequential tasks, familiarity with the use of algorithms (e.g google maps) and more objective tasks in general.

Question 4

Why is self driving AI seen as a good thing according to the lecturer?

Answer 4

It is predicted to be able to prevent up to 90% of traffic accidents. Automated cars do not suffer from stress, tiredness or get drunk etc.