Why the jury is still out in the case for robot-powered PR
The chatter around artificial intelligence has grown a lot louder over the last 12 months since the launch of ChatGPT, the all-singing and dancing robot that’s allegedly coming to take our jobs.
A new report by management consultants McKinsey claims that up to 70% of the work that many of us do could soon be automated. Just this month, a Court of Appeal judge admitted asking the “jolly useful” ChatGPT for help in writing a judgment and being satisfied with the answer.
The headlines didn’t tell the full story. Lord Justice Birss did use ChatGPT, but only to summarise an area of law in a paragraph of a judgment otherwise written by him. Furthermore, as an area of law he knew well, it was a question he already knew the answer to, and he took pains to stress that he still took full personal responsibility for the content.
The very fact that it made the news shows that it was viewed as an experiment and, while successful, Lord Birss did not indicate that he and his colleagues would be hanging up their robes just yet.
With such high-profile endorsements, however, AI is no longer a curiosity we can afford to ignore, but neither is it a finished product or a like-for-like replacement for PR.
As legal PR experts, accuracy is paramount. But good PR is also about persuasion. It engages, it is knowledgeable yet nuanced, it understands and speaks to its specific audience with empathy. It is not churned out in a way that businesses used to produce content to improve their SEO rankings, packing it with as many relevant keywords as possible but otherwise adding little value.
Google’s priorities have changed to be more in line with what users want, favouring “original, helpful content written by people, for people”, and that’s where well-crafted content comes into its own.
ChatGPT is not basic, far from it. The reason that the new(ish) kid on the block has blown competitors such as Meta’s ill-fated Galactica out of the water is that it is already so advanced. The brains behind the bot of the moment, OpenAI, apparently hired hundreds of people in the months before its launch to effectively train it in how to better respond to questions from humans.
Development is moving so fast that even in the last few days, OpenAI have announced that ChatGPT will soon have the ability to hold conversations like Siri, Amazon’s Alexa or Google voice assistant and interact using images in a similar way to Alphabet’s Google Lens.
It will, OpenAI says, be able to tell bedtime stories, settle debates at the dinner table, and even help you decide what’s for dinner by analysing a photograph of the contents of your fridge. For PR, however, it is at the moment a useful aid to information gathering rather than a tool to produce the ready-to-publish article.
It will inevitably evolve and become more advanced but, like other technological developments that have come before it, it will still need some degree of human input to review, refine and maximise its effectiveness. That is likely to mean not only a writer’s touch to ensure engagement, but also to check it for sense, accuracy and sensitivity before publication. Any complaints that the content isn’t accurate, appropriate or performing as well as it should are likely to come to you, and there isn’t a bot you can pass the buck to for that just yet.