As CES kicks off in Las Vegas this week, and the tech-bros showcase a flurry of innovations, the ad industry will no doubt be finding out what AI can do next. Affinity’s executive creative director Matt Batten gives pause to ask a burning question: what’s at risk in pursuit of creative speed and cost-reduction?
Advertising doesn’t build brands in moments. It builds them in layers.
Decades of evidence-based thinking – from Byron Sharp and Ehrenberg-Bass to the broader body of mental availability research – tells us that brands grow by being remembered, recognised, and emotionally reinforced over time.
Through repetition. Through consistency. Through distinctive brand assets that consumers come to trust almost without thinking.
Logos. Characters. Colour palettes. Sonic cues. Tone. Casting. Humour. Humanity.
These aren’t decorative flourishes. They are memory structures. They are how consumers feel about a brand before they consciously decide anything at all. And embedded within many of those structures has been a quietly powerful value: authenticity, the sense that a brand means what it says, and says what it means.
AI hasn’t disrupted advertising’s tools. It’s disrupted those structures. By the time you can spot the trick, the magic is already gone.
That distinction matters because this isn’t an argument against AI.
Used properly, AI is already proving its value inside agencies and marketing teams. It has become a powerful tool for exploration and acceleration: proof of concepts, visual stimulus, prototyping, storyboarding, placeholder visuals and audio, and rapid exploration of visual styles. In these contexts, AI functions like a sketchpad. Rough, provisional, and understood for what it is by experienced professionals.
The danger emerges when that internal tool becomes a public-facing shortcut.
There is a fundamental difference between using AI to explore an idea, and using AI as the idea, especially when the output is pushed directly to consumers, at scale, under the weight of brands that have spent decades building trust and emotional meaning.
The Illusion of Effortlessness
Over the past year, industry feeds have been filled with AI-generated “ads” proudly declared as fast, cheap, and frictionless. Many are framed as proof that traditional creative processes are bloated, slow, or obsolete.
But these pieces are rarely born from real briefs. Genuine client briefs are complex. They come with brand codes, legal constraints, commercial objectives, competitive realities, and stakeholder scrutiny. They demand trade-offs. AI demos demand none of that.
Without a client, there’s no approval process. No pushback. No uncomfortable question of why. As a result, many executions lack a core idea or insight. The concept is reverse-engineered in the edit from whatever the model happened to generate. Rather than gathering the right Lego bricks to build a beautiful and functional racecar, it’s akin to being handed a random handful of blocks and seeing what kind of Frankenstein’s car could be cobbled together. Good luck!
Some AI creators have gone further, recreating famous ads frame-for-frame (like Apple’s iconic and groundbreaking ‘1984’ TVC) merely parroting work that took scores of professionals months to create, then trying to take credit for the
AI ‘achievement’. Yeah, it’s easy to look impressive when you’re standing on the shoulders of giants. Congratulations, you’re the newest member of the John Myatt Fan Club. Look him up.
There’s also a ridiculous short-sightedness among these lone operator ‘AI-studios’ bragging how their latest output was done in an afternoon for the price of a decent lunch. How will their business model be profitable after setting such client expectations? Cheap, fast, and indistinguishable isn’t disruption. It’s turbocharging the industry’s longstanding ‘race to the bottom’.
The Bigger They Are, the Harder They Fall
Increasingly, AI-generated ads are moving beyond the purely speculative territory. Big brands are now deploying fully AI-generated advertising, and discovering that audiences are far less forgiving when real equity is at stake.
Recently, famously, and embarrassingly, McDonald’s pulled an AI-generated Christmas ad after intense backlash. In a stark misalignment for a brand built on warmth and shared moments, viewers described it as “cold” and “emotionally hollow”. Ouch.
Valentino faced criticism after consumers labelled its AI imagery “disturbing,” an uncomfortable response for a luxury house synonymous with human craft.
Coca-Cola ignited widespread backlash with an AI-driven Christmas execution of their iconic and much-loved trucks, which consumers described as “soulless,” “creepy,” and “creative bankruptcy.” When a brand famous for emotional, human-driven storytelling struggles to make AI feel human, the warning is clear.
Locally, Amaysim’s AI-led work split opinion with praise for experimentation countered by concerns around authenticity and obvious shortcutting. Not to mention the waxy, uncomfortable, and disturbing aesthetic. When consumers are so focused on looking for a six-fingered hand, are they really paying attention to your brand (in a good way)?
An unofficial AI-generated ad for Liquid Death is often cited as a success case. It went viral and its absurdity felt on-brand, but Kantar testing revealed poor brand recall, falling in the bottom 25% of all ads tested by Kantar. The spectacle was remembered, the brand less so.
In fact, Kantar partnered with Affectiva to combine survey data with cutting-edge facial coding to capture genuine, intuitive reactions of people watching AI-generated ads. Their results showed that while gen-AI ads evoke surprise and smile as a first look, they report stronger negative emotional reactions, over-indexing for confusion, sadness, and ‘brow furrow’. Furthermore, the study found AI ads scored below average for branding, especially when AI’s involvement is obvious.
Which also means there may actually be a safer AI approach.
Leaning into characterisations and animated styles rather than attempting human realism could be an acceptable way to distract consumers from reeling at AI-slop and waxy people with soulless eyes.
For example, Google’s AI ad for Thanksgiving avoided the pitfalls of uncanny valley and any potential backlash by blending an adorable felt-like AI turkey in a stop-motion style with some post-production animated frames.
The pattern is consistent. When AI attempts to replicate human nuance and emotion, audiences recoil. But when it’s framed as deliberate artifice or modernised animation, tolerance increases.
Laughing With You, or At You
Today’s audiences are highly visually literate. They spot AI instantly — the strange physics, unnatural lighting, illogical movement, and deadened eyes. Once they clock it, the conversation shifts. The work stops being judged on what it says and starts being judged on how it was made.
That’s where backlash lives.
If a $4,000 handbag is sold with visuals that look like they took four minutes to render, the disconnect is immediate. AI doesn’t just risk making ads look cheap. It risks making brands look careless.
Backlash Is a Media Problem Too
The risk isn’t only reputational. It’s economic.
As per AFFINITY’s War on [Media] Waste, we’ve already collated the research from 50+ industry sources to show an estimated 20–40% of media spend is wasted — lost to poor attention, mis-targeting, low-quality environments, or ineffective creative.
Now consider AI backlash on top of that.
When an AI-generated ad is panned and pulled, the spend behind it is unrecoverable. Media booked and paid for cannot be clawed back. There are no makegoods for lost goodwill, no optimisation levers for disappointment.
In an era of intense scrutiny on marketing effectiveness, an AI backlash doesn’t just underperform. It becomes visible evidence of poor judgement.
The decision to save on production costs can become very costly indeed.
The Copyright Blind Spot
There’s another risk quietly emerging: ownership.
In many markets, copyright protection applies only to work created by humans. Fully AI-generated assets may sit outside protection entirely. That means expensive, high-profile creative can be legally repurposed by others.
We’ve already seen this play out. After McDonald’s pulled its AI ad, an indie challenger agency in Pasadena – All Trades Co. – reused one of the ad’s AI-generated characters to mock Maccas in the agency’s own ad that reinforces their human-led positioning.
Legally ambiguous. Culturally brutal.
Now imagine if Hungry Jack’s had done it as a symbolic message of siding with the consumer’s outrage in their ever-continuous burger battle.
Innovation or Shortcut?
AI can absolutely amplify creativity. Used well, it accelerates exploration and expands possibility. Used poorly, it exposes the absence of an idea.
So far, the most damaging AI campaigns haven’t been offensive. They’ve been empty. They felt like brands chasing novelty instead of meaning.
If you’re going to use AI, own it. Take the creative opportunity to lean into it, intentionally, obviously. Make it unmistakably deliberate. Make it serve an idea rather than substitute for one.
The Real Risk
The real danger isn’t one bad ad. It’s erosion.
Erosion of trust. Erosion of distinctiveness. Erosion of the belief that a brand cares enough to get it right.
Once consumers sense corner-cutting, that assumption spreads — into perceptions of product quality, service, intent, and support for a workforce made of the same humans to whom the brand wants to sell.
AI will keep evolving. Norms will stabilise.
But the fundamentals won’t change.
If your advertising looks disposable, consumers will assume your brand is too.
And that’s a cost no efficiency gain can offset.
