Too late to close the AI stable door?

Artificial intelligence (AI) is all over the news at the moment, and well it should be. As a journalist, I’ve seen what Chat GPT can do and it’s both wonderful and terrifying. 

Actually, I’m not too bothered at the moment about the threat it poses, because:

1) It’s a couple of years behind with information (no good for news!)

2) It can’t properly discriminate and sift through bias and falsehood, and

3) It can’t replicate the tone and style of a good writer, and draw on their experiences.

4) Oh yes, it’s also a bit clunky.

So it’s not perfect. Will it improve? Yes. The version we have now will be refined and modified and I’ve no doubt will prove a serious challenge when it comes to churning out ‘basic’ copy.

What about AI when it comes to underwriting, I hear you ask? Well, again, it has come on in. leaps and bounds but we are not quite there yet. The recent pricing debacle in Australia has demonstrated that an algorithm can sometimes get it wrong, with serious consequences for the wider market.

Still, I’ve no doubt that AI underwriting will improve with the years, and that underwriters should be wary about where all of this is going. As a senior market figure gloomily said to me recently, there will probably come a time when AI will be able to underwrite even the most complex risks in the market.

How do we react to this? Quite frankly, I don’t know. But I really think that all of us professionals need to have a long, hard dialogue with the tech sector, government, and most importantly, ourselves. Where is all of this heading and are we comfortable with it? The horse may have bolted, but we can try and tame it, and make sure we have a more comfortable ride, at least.

Marcus Alcock,

Editor, Emerging Risks