This week, I added “ChatGPT” to my Microsoft Word dictionary, as the fabled word processor duly appended its red squiggle underneath the term. I’d urge you to do the same when the opportunity presents itself, as this is a tool that isn’t going away any time soon.
If you have missed it, ChatGPT (Chat Generative Pre-trained Transformer) is an AI-based chatbot, accessible via a web interface, that can respond to queries based on its reinforced learning model. Essentially, you can ask it questions on a variety of topics and it can prepare responses of a quality significant enough that academics are concerned over its use.
Indeed, it has been reported this week that Google are looking to reinforce their position as search leader following Microsoft’s significant investment into OpenAI, the team behind ChatGPT. And that’s probably the most significant element here; it’s not what ChatGPT can produce, it’s why people might wish to use it.
The evolution of internet search has been leading us in this direction for years. In the early days, you’d type in a keyword and where that word matched a database, you’d get a hit. Then Google started crawling the web and automatically building databases, and after that it evolved into ranking and scoring, and today, you can ask your search engine questions to find a multitude of similar queries from others. How people use the returned links updates the search algorithm, with the significant results promoted, to an extent that most of the top 3 to 5 results – outside adverts – are going to be of some use to you. Rarely do we venture on to page 2 or 3 of Google any more.

Yet this success in search has had subtle changes on how we look for and curate answers. I was at a technology summit a couple of years back where search was being discussed, and a panellist said that consumers are becoming more demand-driven. They just want the answer and solution to things, not have to find content related it to figure out by themselves.
This is the real significance of ChatGPT. It changes the way we curate responses. We might not need an essay written for a University assignment, but we may want step-by-step instructions to repair a household item. We might want it to write some code for us to complete a piece of software we’re working on or write lyrics for a catchy new tune. It can blend our nascent search for answers with inspiration and creativity.
There are problems, of course, and even OpenAI admit that they haven’t got it perfect just yet. The ambiguitity, bias and fact-checking limitations of ChatGPT’s output means it should be used with a healthy dose of disclaimers. But the tool is version numbered – a clever, maybe subscious marketing trick – and it will improve over time. That’s how reinforced learning works.
Microsoft expects to deepen its relationship with OpenAI by offering access to ChatGPT within its Office suite. Think about that for a moment. How significant could that be in the workplace? Virtually all human work has some element of problem solving and research to it, and having an assistant on tap (no, not that one!) to automate this could change how we use search forever.
Leave a Reply