The AI Arms Race Nobody Asked For: Why Marketing’s Future Might Be Bureaucratic, Not Brilliant
A (Mostly) Satirical Look at the Future... But Only Just
This is a story about the future of marketing. And yes, it might sound absurd. But like all good satire, it’s only absurd because we’re speeding towards it faster than anyone wants to admit.
Now I think AI has the potential to have a really positive impact on the advertising and marketing industry. If done well it will improve the effectiveness of our outputs, make the industry more inclusive and help with some of the challenges we are currently facing.
However, in our natural drive to stay ahead of each other, we tend to prioritise efficiency over effectiveness, rapid demonstrable results over long term success. This isn’t just a marketing issue. As Nobel laureate Herbert A. Simon warned decades ago, a wealth of information can create a poverty of attention, systems optimised for speed and scale often end up brittle, blinkered, or just plain wrong. We optimise what we can easily measure, rather than what truly matters. And so, rather than redesigning marketing systems for what they could become, we simply reinforce and refine the logic of what already exists.
I fundamentally believe that our job as marketers and advertisers is to control a closed system, but to work in uncertainty, stacking the odds, navigating complexity, and nudging probability in the right direction, not for intellectual curiosity or artistic integrity, but to deliver significant sustainable commercial value. That’s what makes marketing challenging, and rewarding. But the current trajectory seems to double down on a fantasy of perfect control: endlessly optimising systems built for measurability rather than meaning. The result? AI systems arguing with each other in endless, beautifully formatted PowerPoints, while humans, once known as strategists, become little more than referees in a bot-on-bot death match.
So, let’s take a little trip into the world we’re building — and ask whether it’s one we actually want to live in.
Imagine This: A Marketing Meeting in 2028
You’re sitting at the head of the boardroom table, tasked with signing off a media plan. But gone are the strategists and planners you used to spar with — the ones who would challenge your assumptions, push for deeper understanding of consumer needs, argue passionately for a big creative bet, or question whether TikTok really needs another branded dance.
Instead, you’re facing very expensive, very serious AI-generated dashboards.
The agency’s AI presents a sleek, dazzlingly complex plan, backed up with predictive analytics claiming perfectly specific outcomes that could, frankly, be a market leading quantum computing or a Harry Potter spell for all you know.
The auditor and consultant AI(s) audibly sigh, despite being an electronic dashboard, then dryly note that unless you move 20% of spend to linear TV, it forecasts a “likely underperformance against benchmarks.”
Then four separate Platform AIs dial into the meeting, each demanding 80% of the budget, promising that if their recommendation is followed, your brand will not only double in size but also solve climate change, win a Cannes Lion, and achieve sentience.
Your own in-house AI? It’s flagged all the plans as "sub-optimal," and after a few whirrs of synthetic thought, calmly recommends cutting the comms budget by 90% to improve ROI by 300%, and reallocating the rest to ‘interactive holographic avatars in the metaverse's third-favourite digital food court.’
And so, there you are; expected to make a decision between artificial minds that fundamentally disagree, while your own (very human) mind quietly screams into the void.
This isn’t a dystopian sci-fi thought experiment. It’s a direction we are already heading toward.
The Promise of AI: Faster, Smarter, Frictionless... Allegedly
AI promised to make marketing better. That was the pitch. Faster insights, smarter targeting, cheaper optimisation, less admin; delivered without human flaws, friction, or hangovers.
It’s a compelling promise. And in theory, AI could deliver on much of it:
Instant audience selection, with a side of overconfidence.
Media mix models that update faster than your strategy can keep up.
Creative tested, placed, re-tested and forgotten all before your coffee's cooled.
What’s not to like?
But here’s the catch: we are all building our own AI; each with different logic, different assumptions, and zero interest in cooperation.
Agencies have AIs to plan media. Auditors and consultants now have AIs to verify them, like a particularly suspicious older sibling checking if you've really done your homework. Platforms are running activity with their own AIs. And clients, naturally, are building their own AIs to decode all of them and integrate into the wider business.
So instead of replacing human error, we are creating a battlefield of bots, each claiming to be right, each grounded in opaque assumptions; the kind so impenetrably complex you'd need a PhD, a decoder ring, and a séance with Alan Turing just to make sense of them, each throwing out charts, graphs, and lovely-looking dashboards that don’t agree.
What AI promised to eliminate, confusion, complexity, and contradiction, it may soon industrialise.
Or as the Vogons from Hitchhiker’s Guide to the Galaxy might say:
“Resistance is useless. Here is Form B-27-AI-42, in triplicate, to file your disagreement. Please return copies to Departments 7G, 12Q, and 19Z-A through the pneumatic messaging system, which is currently undergoing unexpected maintenance.”
Enter the AI arms race: agencies, clients, auditors, consultants, platforms and an ever expanding list of new 3rd parties all racing to develop their own proprietary AI systems, each one claiming superior insight.
The New Bureaucracy of Bots: From Creative to Diplomatic Corps
And here’s the punchline: when AIs fight, it’s going to be humans who have to pick up the pieces. Enter a new kind of role: AI mediator. The person who has to explain why one AI says X while another says Y. The reluctant diplomat trying to broker peace between warring algorithms. Talent won’t be freed up by AI — they’ll be redeployed as 'AI whisperers', tasked with justifying conflicting machine logic in rooms where strategy used to live.
As Tom Goodwin rightly observed, "What if AI doesn’t remove bureaucracy; what if it adds to it?" The cost of all this? It’s not just time and money, it’s our craft, judgement and sanity.
When every decision becomes an AI negotiation, human creativity and strategic thinking get side-lined:
Craft gets crushed under standardised AI outputs.
Judgement eroded as people defer to "what the AI said."
Innovation stifled if deviating from AI "rules" is seen as risky or unjustifiable.
Mad creativity and absurdist problem-solving disappear from the equation; if AI can’t even generate a picture of a full wine glass, where will the next sideways idea come from?
Instead of making marketing smarter, we risk making it dumber, compounding the challenges of today with a technocracy where algorithms dictate, and humans defend.
Designing the AI-Supported Marketing System We Actually Want
Of course, there’s nothing wrong with using AI to make marketing more efficient. But if we’re going to rebuild our tools and systems for an AI-enabled world, let’s not end up with a shinier version of the old one. Let’s design for the system we actually want — one that uses AI to elevate decision-making, not drown it in dashboards.
First, put humans back at the centre.
AI should support human judgement, not replace it. It should offer options, not orders. As Taiichi Ohno, creator of the Toyota Production System, advocated with his 'Five Whys' technique, asking 'why' five times to unlock the root cause, our AI systems should be designed to support marketers in doing exactly that. Not to close down discussion, but to provoke it. To nudge us beyond the superficial, prevent distractions and help us get to better, more human outcomes.
Second, demand transparency of logic.
We have to accept that we’re fast approaching a point where AI, and possibly quantum computing, can make connections and calculations that the human brain simply can’t. That’s not a threat, it’s a reality. But we must still understand the flow of reasoning, the steps taken to get to an output. That kind of structured transparency doesn’t diminish human input; it elevates it. It gives marketers the tools to interpret, challenge, and contribute to better decisions, rather than just nodding at the dashboard.
Third, let’s stop building single-use AIs for the sake of it.
The real value won’t come from deeper silos, it will come from AIs that bridge disciplines, connect use cases, and help teams think across strategy, media, creative, and measurement. Unless your AI genuinely adds or integrates new thinking, rather than just rehashing public data in a fancier font, perhaps we don’t need one more bot in the ring.
And fourth, ask better questions before you buy.
Too many AI tools are being adopted with a single question in mind: 'How much headcount can this replace?' or 'How much efficiency can it deliver in the next quarter?' That might offer short-term gain, but it leads to long-term pain. We risk building an industry optimised for cost-cutting instead of value creation, one where we hollow out human capability in the pursuit of dashboard-friendly margins. Every marketer should be able to answer the following before bringing a new system into the fold:
Who sets the AI’s assumptions — and can I see them?
How does this AI handle conflicts with other AIs?
What role do humans play in checking or challenging AI outputs?
Can I explain this AI’s decision to my CFO?
What happens when it gets it wrong?
We still have a choice. Not about whether AI will be part of marketing, it already is,
but about the kind of relationship we build with it. Do we want a future where marketers are curators of intelligent systems and advocates for creative possibility? Or just referees in an endless bot-on-bot shouting match? Now’s the time to decide.
Because I really do not think marketing doesn’t need more machines to argue. It needs more space for humans to think, to create, and to lead. Achieved not in spite of AI, but because of it.