Ethics at the Speed of AI with Prof. Aini Suzana Ariffin & Liz Longworth | Aotearoa AI Summit Special

So with the framework, because there's this concept of unintended side effects.

And actually fundamentally unintended side effects are one of the biggest risks with frameworks because when you say something like do no harm, that can absolutely be applied in a particular context.

But if you haven't fully anticipated how the adjacent or subsequent causality of you adhering to the do no harm policy or rule applies, you can actually create a new risk that the framework doesn't actually handle.

So it'd be really interesting to understand what parts of the unintended side effects and what provisions there are for revision as we learn, because one of the areas in technology that we completely missed was social media.

There was no governance frameworks.

It sort of slipped in as another website

and society sort of accepted it and it grew

and grew and grew into a bit of a monster.

And it changed the world, both good and bad.

It made people more connected, but the unintended side effect of that connectedness caused a whole bunch of health issues.

And now we're starting to see repercussions of that very late in the piece, right? So I think AI has a lot of those risks attached to it.

And it'd be really good to understand why adoption of a framework like this is a good thing.

You know, you've explained that really well, but how does the framework also handle potential unintended side

effects.

While you're reflecting on that, I should say that these things are not fixed in any one time.

So if you adopt a principles-based approach in your guidance, it enables you, as the technology evolves and goes down a certain path, it maybe starts to really run a butt up against um the the human dignity of the human values or embeds biases that we don't want whatever it might be um you can course correct because you have this sort of very it's both flexible but it also provides a clear expectation about what the value value should be and embedded in the way the technology operates and is deployed.

So I think that provides some comfort.

And I should also mention that it's not binary in the sense that one of the big themes from this conference, and coming back to your question Hamish about why we're here, Susanna was here to participate in the concept of how you build trust and social licence and I was astounded to see that all the business speakers, everybody who's taken the stage has started from the presumption that that in order to be successful, in order to have the economic outcomes that they're looking for, you can't do it without having built the trust.

And that requires the ethical guardrails.

And you're seeing that just in our ordinary citizens' debates now.

So the two just...

the two instead of being over here and over here the two are actually integrated now the concept of uh trust um mechanisms to build trust to get to to get the life the social license and to continue they're not either one or the other but uh

Because understanding the principle is very crucial.

So what we did, you can have it in the Buddha, we come out with very, what you call it, easy understanding of each of the concepts.

So we come out with the brief explanation, what you mean by the not harmful.