A hot topic at the moment in PR, and many other industries, is AI – artificial intelligence. People in all sorts of businesses are wondering how this ground-breaking technology will affect the direction of their industry and, ultimately, employment.
The truth is, at the moment, it is very difficult to tell. I do think a lot is being made of some, on the face of it, very impressive results. I’ve heard of scripts that fairly accurately replicate popular sitcoms and songs that seem to follow the style of particular songwriters. On closer inspection, though, I’m inclined to follow Australian songwriter Nick Cave’s response that actually what is produced often “sucks.” His argument is that AI and the results it produces are in fact a “grotesque mockery of what it is to be human.”
I think there is some truth in what he says, but that won’t be the end of it. However much we might want to, we can’t escape the arrival of workable AI. In March 2023, over 1,000 experts in the field (and Elon Musk) published an open letter calling for a six-month moratorium on AI development to allow us to work out how we are going to deal with the repercussions of this new technology. So, it obviously is a concern. The fact is, AI is here and we, as a society, have to learn how to deal with it.
I’m reminded of a discussion I heard about social media that referenced the arrival of new information technology. From the printing press onwards, all new technologies have been met by founded and unfounded concerns over their ability to destroy society. The issue has always been, not the technology but our ability to control it. As it was expressed in the discussion, it’s not the technology but the user manual we have.
This is certainly true of everything up to and including social media. The systems are not inherently ‘evil’ or ‘bad’, but the way we manipulate them can be. But is the same true of AI?
It won’t come as a shock to find out I’m not an expert on AI. I’ve listened to a few interesting expert discussions (as an entry point, I would suggest an edition of the Rest is Politics podcast issued on 31st May 2023) and had a bit of a play with some of the online AI systems to see what they can do. In terms of PR, though, I think one thing is very clear - AI isn’t the answer.
Let me explain. Essentially public relations is the art of communication. It is the ability to build a connection between a client and potential customers. This is, at its very core, a human connection. I’ve written before about how ‘people buy people’, they don’t buy automation. Just think about our response to automated phone systems – you put up with it, but you don’t like it.
It would be naïve to think AI won’t affect PR. There are simple tasks that a computer, especially one with AI, can do faster and more efficiently than a human. It can sift through information and create a synthesis far quicker than any of us. This is a real advantage and it will certainly have some amazing real-life applications, but what is produced cannot necessarily be described as ‘good’.
To put it simply, without human oversight, would you be happy to just let AI create your marketing? We must remember, there are already documented instances of AI bots creating very impressive-looking documents with references to other texts. It is only upon further inspection that it becomes clear these references are fabricated by the AI system.
There’s also the issue of what information the AI system is using to create its text. As a trial, I recently went on to the ChatGPT system and asked it to create 200 words on a company I know very well. At first glance, the results seemed impressive. It did write exactly 200 words on the subject and the information contained in the text was not wrong.
However, it also wasn’t exactly 100% right. For example, this is a global company that is constantly acquiring and amalgamating facilities around the world. Its statistics – employees and number of buildings – are constantly changing, and these are always updated on its website and every footer used in online communications. In the time I’ve worked with the company, the figure for offices and laboratories has shifted between 2,650 to 2,950. The current figure is 2,750.
The Ai system has obviously gone and looked online for data and this was one of the statistics it found but, because the number is constantly changing, it couldn’t decide on the correct figure and so settled on “more than 2,600.” Not incorrect, but also not correct. Since this is a company whose reputation is built on accuracy, not being correct is a real problem.
The second issue is that, while the text exactly hit 200 words, it wasn’t exactly engaging. In fact, I was reminded of Nick Cave’s, “grotesque mockery.” What we ended up with was a parody of an answer. The AI bot had seen the request as a problem to solve and had gone to find the answer.
A lot of writing is actually editing. The first stage is often to solve the problem – write 200 words that answer a question – but then the rest of it is making sure it feels more natural and communicates the message more effectively to the reader.
I would say that the AI system wrote the first part very impressively, but it hadn’t done the second part at all. Communicating with other humans is a complex business. In our own lives, we know that what we think and how we express it, even if we express it very well, isn’t necessarily how it will be received and understood by the other party.
If PR is about the art of communication, therefore, it is imperative that humans remain at the forefront of your messaging. You need to be certain that your message is accurate and is being delivered in a way that your customers will appreciate.
Image by Gerd Altmann from Pixabay