Finally people are talking about AI and ethics in the same sentence. Its been a long time coming…It seems that we don’t always like what the algorithms are telling us and we now face
an ethics problem.
A Google search reveals that a popular remedy to reinstate ethics to AI is to focus on a thing called “bias removal”…
It stands to reason that in bias removal we need to look at what we feed these AI algorithms…
Where did you get the data?,
why did you include the data in the algorithm?,
should you even have been collecting the data in the first place?
This then in theory should fix things.
This approach goes against the idea of “more data is better” and “let the algorithm figure it out”. We are now suggesting an intellectual cleansing of sorts which should sanitize the outputs.
The problem is that true ethics might well be a human trait and as such will vary based on many things including the context and the culture where it is formed. What might be seen as ethically sound on one side of the planet might well create outcry on the other. That’s the reality of our diverse and wonderful world.
Ethical and moral standards have been a challenge from when we first worked out how to communicate with each other all those very many years ago. Right now though, the best use of AI might be how to incorporate diverse views, opinions and attitudes into our digital world.