In Bruegel’s Landscape with the Fall of Icarus, the boy who flew too close to the sun is almost an afterthought. You have to look hard to find him: a pair of white legs thrashing in the green water, just below a merchant ship. The plowman keeps plowing. The shepherd gazes elsewhere. The ship, as W.H. Auden noted in one of the finest poems of the last century, has its own business to attend to and sails on without a backward glance.
Icarus had been warned. His father Daedalus, who built the wings from feathers and beeswax, told him to fly neither too low, where the spray would soak them, nor too high, where the sun would melt the wax. Icarus climbed anyway. The urge to explore, whatever the cost, is older than Ovid, who set the story down two thousand years ago, and it is no weaker today. We all know we are flying close to the sun on wings of feathers and wax. We cannot help ourselves.
I was reminded of this at a recent gathering in Dubai, where the grandmasters of the new dark arts debated whether artificial intelligence spells the end of the communications industry or offers it a new path. If a conclusion was reached, I missed it. Since everyone in the room had a stake, whether building the models, using them, or defending agencies that now describe AI as a useful tool, perhaps none was to be expected.
What struck me more was happening below the podium. As each speaker stood up, a swarm of photographers jostled for the shot, every one of them armed with a camera that does the dull work of photography for them. Focus, exposure, light: solved. A shot that once took a Cartier-Bresson or a Man Ray a minute of careful composition is now available to anyone who can point. Not one speaker mentioned it. Yet there, perhaps, was the answer. When the camera does the craft, the value moves to knowing where to stand and which moment matters.
Imagine Bruegel’s canvas repainted today. The ploughman is gone, replaced by a GPS-guided tractor. A drone herds the sheep. The sailing ship is a supertanker or a Type 26 frigate, and Icarus is flying a Raptor. Every figure in the picture has been automated out of it. Yet we have survived and prospered. The question is whether communications will do the same.
Start with what AI has already mastered. The execution layer is vanishing: copy variants, first drafts, media plans, social content, production, translation, monitoring, reporting, all are gone or going fast. In advertising, the platforms are cutting out the middleman. Meta’s and Google’s automated campaign tools now write, target and optimize themselves, and Mark Zuckerberg has described a future in which advertisers arrive with a business goal and a budget and leave the rest to the machine. That leaves the holding-company model, large teams billed by the hour to make and place things, structurally exposed. The consolidation and job cuts across the big groups over the past two years are all evidence.
What AI cannot take, at least not yet, is judgment under pressure. In a crisis, a chief executive does not need a statement generator. They need someone who has seen this before, knows which journalist will run with it, and has the nerve to tell them to say less. AI can draft ten options. It cannot take responsibility for choosing the right one.
Nor can it replace access. Public relations trades on trust: with editors, ministers’ offices, regulators and boards. A model has no reputation to stake and no phone number anyone will answer. Clients pay, in part, for accountability: someone to call at two in the morning, and someone to blame. And AI lacks taste. In markets like the Gulf, where protocol and who says what to whom still carry real weight, fluent, competent work is not enough. Today everyone clears the bar, competence is nothing.
The losers are easy to name. Advertising agencies are more exposed than PR firms, because more of their value lay in production and buying. Large agencies are more exposed than boutiques. Junior staff are more exposed than senior counsel, and that is the industry’s real long-term problem: it is sawing off the bottom rungs of the ladder that trained its future seniors.
The winners will be small, senior-heavy advisory practices that use AI for the grunt work and charge for judgement, access and outcomes rather than hours. That looks less like a traditional agency and more like a consigliere with a very fast back office. AI will not replace the adviser. It will replace the agency as a business model, and it is already doing so.
So should anyone start a career in communications today? In the conventional sense, no. Entry means joining at the junior level, precisely the layer AI is hollowing out. Agencies are hiring fewer graduates, entry pay is poor, and the path from copywriter to creative director is losing its lower rungs.
The exception is those repackaging a career they already have. Those who have spent their careers advising ministers or supporting Fortune 500 chief executives, and that profile (senior counsel, real access, rare writing skill) is the one most likely to stay in work while machines handle production. The opportunity lies in a small advisory practice, not an open-plan desk at an agency.
For the young, the advice is to arrive with a sharp edge: strategy, data and media, or deep knowledge of a sector such as government, health or luxury, combined with genuine fluency in the tools. Generalist creative and account-handling roles are the easiest targets for the models to pick off.
Look once more at Bruegel. The ploughman does not stop. The ship sails on. The world was never going to pause for Icarus, and it will not pause for the agencies either. The question is not whether the communications industry falls, but whether those who practice it can learn the lesson Daedalus tried to teach: fly high enough to see, low enough to survive. The wings are the machines. The judgement about altitude is still ours.
(Ian Philion is a strategic adviser and writer who has worked for more than a dozen presidents in the USA, Europe and Africa. He has also advised and written speeches for a wide range of Fortune 500 CEOs)
