Everyone’s talking about AI. AI will take our jobs. AI will save our jobs. AI will bring back Steve Jobs. AI is watching you. Everyone’s watching AI. Everyone’s listening to AI. Did I mention AI? Who am I anyway?

Aside from the semantic question that if AI was literally intelligent then could it really be artificial, there’s rarely a conversation about AI that hasn’t already been had. But unlike most topics, AI fatigue is unlikely to happen any time soon because AI is magnificent at making sure conversation piques your interest. In some fields – communications! – keeping eyeballs is its primary function.

Yeah, so this column is about AI. It comes with an 89% guarantee that it’s probably not been written by AI.

Back to that ‘keeping eyeballs’ thing. I once worked on a spec script where an AI misinterpreted the need to retain customers’ eyeballs and created an android Sweeny Tood/Gunther Von Hagens-type that went around harvesting the peepers of hapless residents of the Lake District. This probably (but not definitely) won’t happen in real life, but like all Black Mirror rip-offs the preposterous had at its centre something pretty prosaic. Currently, AI is only as good as the input, just ask the legal team at Air Canada. AI is not self-driving (but it soon will be).

There are thousands of articles – some actually written by humans – that predict the future of AI and the profound impact it will have on communications. We must all become, we have learned, conductors rather than the orchestra. And that’s great, because to be honest I’m rubbish at playing the bassoon and my timpani drum leaves much to be desired.

As we conduct our orchestra, we perfect messaging, channels and audience, until our communications is so pure and refined that De Beers makes a bid for it.

But I worry that there’s a huge assumption being made.

Whether it’s a chatbot, an automated ad campaign, or a press release, we define an audience and assume that it is humans that will receive/interact with whatever it is we send out into the world. But already, that’s not always the case. If AI is delivering a message, then it stands to reason that AI could be reading it. Wade into the festering cesspit that is X (formerly Twitter) and read the replies of any large account – it’s at least half bots or engagement farming; click on those and their replies are also automated accounts nurdling around in Elon’s weeds looking for opportunities to engage. The bots are chatting with bots.

The clever AI-generated comms that we put out in the future will very likely read by AI before it hits the few eyeballs in the Lake District that haven’t already been farmed. AI will allow us to become the conductor, but for our targets, it also becomes the filter.

And maybe it never hits a retina at all. Maybe the future is bots that have been ordered to engage, happily engaging with bots that have been ordered to receive. Like Wall-E, forever cleaning-up rubbish because that’s what he was told to do, never stopping because no-one to him to do that either. Round like a circle in a spiral, like a wheel within a wheel, never ending or beginning, on an ever spinning reel…

If you're interested in learning more about AI and its implications for PR, take a look at PRmoment's latest Masterclass: PR & AI. Whether you’ve already begun your professional AI journey or if you're looking for advice on where to start - this Masterclass will help you. At this one event, 12 experts will give 10 Masterclasses on PR & AI.

This PR Observations column was written by David Quainton, head of communications at the digital consultancy Emergn. The opinions are his own.