What is ChatGPT, and should we be worried?

At the end of last year, the rumblings around a new chatbot appeared on the internet and then quickly went viral. Launched by Open AI, ChatGPT (short for Chat Generative Pre-Trained Transformer) racked up one million users in five days (for context it took Twitter nearly two years to hit the 1m mark).

The system is trained on an incomprehensively vast dataset of websites, articles and books – an estimated 175 billion data points – and had users flocking to the site to engage with the potentially revolutionary chatbot and ask it questions such as “How Do I Cook Lasagne”.

However, it was quickly realised that its capabilities were far more sophisticated, and that it could be used to write code, build websites, work as a customer service bot and as a virtual assistant. It can also write essays, copy, and press releases if given a brief.

It was at this point that the questions were posed: is there a malevolent undercurrent to ChatGPT – is it going to steal our jobs, bring the education system to its knees and ultimately usher in a new dawn of communication, creativity, and life as we know it?

Probably not, no.

The main concern for brands, as well as PR and Marcomms professionals and digital media in general, is the fact that ChatGPT is a remarkably efficient search engine (though it’s very important to caveat that its current knowledge ends in 2021; it lives in a world where the Queen is still alive, and Liz Truss was never Prime Minister).

While Googling something will take you to a page of websites and links, ChatGPT will give a detailed and insightful answer, with no commercial interests or bias. You can’t leverage your brand appearing higher up than a competitor, something which is bad news to the marketing and PR specialists who have spent many years carefully curating their online shop windows with content, adverts, and media coverage to give themselves a digital edge.

Google recently issued a “Code Red” in response to the launch of ChatGPT and the threat it may pose to their online advertising revenue, triggering what could be an AI “Arms Race”. Google have announced today that they are launching their own rival, called Bard.

In terms of the day-to-day, many are reporting that they are already utilising ChatGPT in their working lives, generally for simple data analysis and helping write copy, from emails to reports.

However, it is difficult to get anything but a first draft out of the technology. ChatGPT is drawing on data and algorithms fed to it to mimic human responses, but with any mimic, there is a sense of the uncanny about it.

While eloquent, there is a robotic-quality to the copy generated, a clipped over-professionalism, leaving the reader inevitably feeling unsettled if they assumed it had been written by a real human – even if they can’t put their finger on why.

You also can’t build professional and personal relationships with AI, no matter how much they may be able to sound like us, and they are of course not able to offer an original, or creative, thought or idea.

For those who are saying that ChatGPT marks the future, there are just as many saying that it will be relegated to the pile of things that could have changed the face of the world but just didn’t take off, that currently includes Google Glass and the Segway.

What will be interesting to keep an eye on is how ChatGPT will inevitably need a revenue stream to sustain itself with so many users. A subscription model which promises faster response times and access to new features is being piloted, but at the low price of $20 per month, this won’t generate much income for its founders OpenAI. Until then, we must assume that the adage of “if the service is free to the user, then the user is the product” applies, though it remains to be seen to what end.