Teaching Ethics to AI

Recognise the picture? The Terminator!

Oh yes, that seminal film, The Terminator, where Arnold Schwarzenegger stars as the sentient machine sent back in time to ensure humanity's destruction!

It's actually one of my favourite films and the first film to have a sequel that was actually just as good as the first, if not better.

There has been some discussions lately about the advances in Artificial Intelligence and Machine Learning. Some people see this as the greatest advancement in technology that can help to save us from our self-inflicted climate catastrophe, our healthcare needs or the pathway to a more productive life.  Some people, however, are comparing it to the movie "Terminator" or Isaac Asimov's "I Robot".

In fact Max Clifford, adviser to the UK's Prime Minister, Richi Sunak, declared that

"AI systems could 'kill many humans' in two years without regulation"

Many of us think that it is far too late for regulation and, in any case, not all countries or states get along well enough to trust each other.

Hysterical and Sensational

In reality, all that AI such as ChatGPT and Bard can do is to provide an ingenious search function.  The main problem is that it will make some jobs obsolete. There's no real intelligence behind it, it is more like a parrot with an extremely big memory!
Some people, whose jobs are in danger, would say that it is "easy for you to say that".

But is it really just a glorified parrot? Just what will this glorified parrot be able to do in a few years' time?  It has already come so far, so very quickly.

Without thinking about it deeply, the first worry many people may have is of the "big red button": giving AI control over military defence or worse still our adversaries getting there first. Are we already in an AI arms race? is the unvoiced worry.

Well, let's not get into a panic. I think that the most pressing issue is how AI can come to skew our outlook on life. Just look at how affectively it can target our buying intension and some would claim also our voting intension.  Let's face it, if you start to "like" pictures of llamas, how many more llamas will you start to see all over the internet?

So what can we do?

Well, maybe it is not out of our control after all.

AI parses huge chunks of our data and much of our data is the things that we share on social media. It takes that information and regurgitates it and from this it learns our likes and dislikes - how else does it serve you up with more of the same?

We need to start populating that data with something worthwhile for it to learn from; start teaching it empathy, manners and intelligent conversation.

We need to start teaching AI about ethics.

It is for us to teach it; let's start now! Be a good parent to the AI, teach it how to look after us with care and sympathy.

We would do well by teaching it how to distinguish fact from fiction, right from wrong in a compassionate way.

Teach it to save us from the worst of humanity.