Sponsored Article
Is generative AI a threat to PR agencies?
Author : Richard Stone
01 February 2023
The popularity of ChatGPT since its launch last November has made generative AI one of the hottest topics in tech over the last couple of weeks. It’s been the talk of Davos, the subject of 4,520,000 Google News search results and even the reason a woman in south London left her husband.
But what does it mean for industries where the production of written content forms part of the service offering? These include technical PR, of course, but also journalism and publishing, SEO, copywriting, and market research as well as plenty of others.
Quality or quantity?
My experience of generative AI applications so far, including Chat GPT, is that the copy they produce is normally okay. But okay usually isn’t good enough.
An article produced using generative AI, will normally be grammatically solid and it will read well. But crucially, it will lack passion, supporting references to prove its argument, and originality.
All three of these things are essential if your objective is to create a journalistically valid piece of writing.
Nicholas Diakopoulos, who is an Associate Professor of communication studies and computer science at Northwestern University and a regular and interesting blogger, argued, “It’s probably better to think of these tools as internal newsroom tools, making suggestions to reporters and editors rather than generating text that will be directly published.”
In the context of the current quality of output, he’s absolutely correct and this argument applies equally in PR. Using generative AI to give you headline ideas, skeleton drafts of copy — where you are certain of the facts — or for simple, short pieces of copywriting could be productive.
The drawbacks
The shortfall of the technology is clear, particularly in scientific, technical or technology PR. Sometimes, you will get nonsense. Often you will get content that is technically inaccurate. Nearly always, you will get copy that is as biased as the majority of the content already published on the subject.
Bias, in particular, can be a problem in technical content. Not in the sense of predilection towards or against a particular group, but for or against a particular technique, product, or company.
Imagine that you have launched a technical product that presents a new method of fulfilling an engineering objective. You might have created a way of managing the expansion of a pipe in a building that is superior to a typical expansion joint, for instance.
If all the content that already exists on the subject suggests expansion joints are the way to go, a generative AI will probably create content that argues the same thing. Or, at the very least, you will have to give it so much information in the brief that you may as well have written the piece yourself.
The future
I anticipate three things happening in the world of PR because of Chat GPT and its ilk. The first is that we will all start using it. This is inevitable, so your objective should be to make sure that your organisation uses it well, ethically and with one good eye on its shortcomings.
The second thing is that this could well impact agencies that charge by the hour or day, rather than using a value pricing model as Stone Junction does. If you are a client and you are sent a piece of copy, along with a timesheet that says it took a day to create, you must be confident it actually took a day and that that day was well spent.
If the copy reads like something an AI could have written, which frankly some human-produced copy does, the long-term impact on the agency’s blended day rate will be significantly negative.
In contrast, if your agency is using a value pricing model, you don’t really care how long it took them to write something. You care, instead, about the result; the coverage, the improvements in SERPs position, the website traffic, or the extra sales, leads or other forms of organisational change that come as a result.
All of this means that every agency will have to focus more on its consultancy, its creativity, and the core intellect on which its campaigns are founded.
Finally, I think we could well see legal requirements introduced when copy has been produced by an AI and not by a person. Think about the way an advertorial has to be identified now, in comparison to unmarked editorial content.
Video killed the radio star
One thing is certain, I’m expecting change. The people at the forefront of that change are the ones who will benefit most. And, as with the popularisation of the internet as a news medium, the advent of social media, the introduction of word processing and the killing of the radio star by video, the laggards will suffer.
My feeling is that, if humanity can steer clear of using generative AI as a good reason to leave its husband, we are probably safe. If the PR industry can, at least metaphorically, do the same, the future might well be bright.
Contact Details and Archive...