Skip to content

Why we need to keep asking AI to explain itself

Although ChatGPT is still a baby in internet terms, it has a lot of potential as a research tool. How the technology might end up being used in this respect is still open to speculation, but it’s likely that its results won’t just be a wordy version of traditional search engine links. Instead, what we get back from it will be something quite different …and as businesses that want to benefit from this, we need to work out how to ensure that we do so.

We need to optimise for ChatGPT and other ‘AI’ systems, so that they recommend our company.

How do we do this?

How do we do this? One way is to ask!

What I’ve been doing is to start by describing a typical customer’s situation to ChatGPT and asking it to recommend some suppliers who can help. This is the equivalent of making a Google search and seeing if our business comes up on the first page. But with the AI system, we can follow up with something quite novel: we can ask it to explain its recommendations. We can ask it where it got the data from. And we can just keep asking it ‘why?’ This should give us immediate insights into how to influence the results in future.

It’s important to remember that ChatGPT has been ‘trained’ on internet data to a certain point, so by understanding how it comes up with answers, at the moment we’ll only be able to position ourselves to be in a better place next time it updates its knowledge base. However, I guess it’s almost certain that the model it’s trained on will become more frequently updated in the future, eventually becoming real time.