The AI Content Backlash Is Not Really About AI. It Is About Trust.

Photo by Melyna Valle on Unsplash‍ ‍

I keep getting asked the same question by marketers, and I also see it show up frequently now in the discussion channels I’m part of on WhatsApp and LinkedIn, as well as at small-group meetups.

Is there a consumer backlash against AI-generated brand content?

In my POV, YES.

But the question needs more nuance. The backlash is not simply against AI. It is against content that feels lazy, generic, emotionally empty, or disconnected from the brand that published it. That distinction matters.

I am not anti-AI. In fact, I believe AI will be one of the biggest enablers of marketing over the next decade. But after close to three decades across sales, marketing, technology, regional leadership, and advisory work, I have also learnt that every major technology wave creates the same temptation. We over-apply the tool before we understand the job.

Marketing is especially vulnerable to this because content is visible, measurable, and easy to scale. So naturally, many teams start there. The problem is that content is also where brand trust, emotional connection, and human judgement are most exposed.

AI can help marketers do extraordinary things. It can accelerate research, sharpen insights, improve segmentation, support personalisation, summarise performance, assist with content production, and make teams more execution fluent. But that does not mean every part of marketing should be automated to the same degree.

This is where I think many CMOs and marketers are getting into trouble. They are treating AI adoption as a production problem when, in many cases, it is actually a judgement problem.

Let Me Put My Bias On The Table

When I started AIdeate Solutions, my intent was not to create another AI consultancy that pushes tools, subscriptions, or retainers that clients may not need.

I do not run a platform. I do not have a product I am trying to sell into every marketing workflow. I am not interested in adding to the noise. My value proposition is very clear: In the AI era, I help redesign workflows responsibly, with governance and human judgement in mind.

In this phase of my career, I use my experience across the craft and science of marketing to help fellow marketers navigate difficult moments with greater clarity. Sometimes those conversations are commercial. Many times, they are honest conversations with people who are trying to make sense of the pressure they are under.

And the pressure is real.

CMOs and marketing leaders are being pushed to adopt AI quickly. Boards want productivity. CEOs want efficiency. CFOs want cost discipline. Technology vendors want platform adoption. Agencies are trying to reinvent their delivery models. Teams are worried about their relevance. Consumers are increasingly alert to synthetic content.

Somewhere in the middle of all this sits the marketer, trying to work out what is genuinely useful, what is risky, what is hype, and what will still matter two years from now.

That is the lens through which I am writing this.

Not anti-AI. Not mindlessly pro-AI. Just pragmatic.

The Problem Is Not AI Content. The Problem Is Lazy Content At Scale.

Let us be honest. Bad content existed long before generative AI.

There was plenty of human-made content that failed to connect, convert, differentiate, or deserve the audience’s attention. We have all seen corporate posts, campaign lines, videos, banners, landing pages, and thought leadership pieces that were technically human-made but still painfully forgettable.

AI did not invent mediocrity. It just made mediocrity faster, cheaper, and easier to distribute.

That is why the phrase “AI slop” has become so common. The deeper issue is not that AI was involved. The deeper issue is that too much content now lacks brand truth, emotional resonance, distinctive voice, intent, and accountability.

If content lacks these things, it will fail, whether written by a person or generated by a model.

The difference is that audiences seem to apply a higher level of trust filtering when they know, or even suspect, that content is AI-generated. That is not just a marketing hunch. A growing body of academic work across journalism, creative media, psychology, and platform studies suggests that people often evaluate AI-labelled content less favourably than human-labelled content.

One study by Altay and Gilardi found that people were more sceptical of headlines labelled as AI-generated, even when the headlines were true or human-made. The important insight was not just that people disliked AI. It was that audiences assumed “AI-generated” meant full automation, without enough human editorial judgement.

That is an important warning for brands.

Because brand content is not just information. It is a trust signal.

Every piece of content tells the audience something about what the brand values. Did the brand take care? Did it understand the audience? Did it respect the cultural context? Did it have something useful to say? Or did it simply produce more material because the machine made it easy?

When audiences feel that a brand has taken a shortcut, they notice.

The Thumb Test For Brand Content

Here is a simple test I often use with marketers.

Put your thumb over the brand name. Then ask yourself: would this content still be recognisable as coming from your brand?

If the answer is no, you do not have a content advantage. You have category wallpaper.

AI is very good at producing category wallpaper. It can create fluent, polished, safe, grammatically correct content very quickly. That has value in certain workflows. But fluency is not the same as distinctiveness. A paragraph can be well-written and still say nothing that matters.

A brand does not win by saying what every other brand can say. It wins because it says something true in a way only it can.

That is where human judgement still matters.

A mentor once told me that good content should touch the heart, the mind, or the wallet. The best content can sometimes touch all three, but not every piece needs to. What every piece does need is intent.

Is this content meant to build affinity? Explain a product benefit? Drive conversion? Shift perception? Support discovery? Defend reputation? Reassure a community? Help a customer make a better decision?

If the team cannot answer that question before generating the content, AI will only create more confusion at scale.

This is one of my biggest concerns with how AI is being adopted in marketing. I’m seeing that many teams are starting with the output. More posts. More variants. More videos. More newsletters. More banners. More captions.

But volume was never the same as value.

Consumers Still Value Human Effort, Intention, And Accountability

The trust issue is not limited to journalism. It also shows up in research around creative work.

Bellaiche and colleagues ran controlled experiments comparing reactions to artwork labelled as human-created versus AI-created. Even when participants viewed the same artwork, pieces labelled as human-made were rated higher on dimensions such as liking, beauty, profundity, and financial worth.

That is a powerful finding for anyone in marketing.

It suggests that people do not evaluate creative work solely by its surface. They also care about the perceived effort, intention, and humanity behind it.

Another study by Horton, White, and Iyengar found that the presence of AI art can actually increase appreciation for human creativity. In other words, AI does not simply replace human creative value in the audience’s mind. It may make people more aware of what they value in human-made work.

This matters because brands operate in emotional environments, not just informational ones.

People do not only ask, “Is this content accurate?”

They also ask, “Who made this? Why was it made? Can I trust the intention behind it? Is there a human being accountable for what this brand is saying?”

That is why disclosure alone is not enough. A label that says “made with AI” may be transparent, but it does not automatically create Trust. Trust comes when the audience believes there is human judgement, responsibility, and care behind the work.

The same principle appears in platform environments. Research by Molina and Sundar on AI content moderation showed that people respond differently when AI is perceived as acting alone versus working with visible human collaboration and transparency.

Again, the lesson for marketers is quite straightforward.

AI with human oversight is easier to trust than AI that pretends to be human or operates without visible accountability.

The HAR Framework: A Better Way To Decide Where AI Belongs

This is why I use the HAR Framework in my advisory work.

HAR stands for Human Agent Ratio.

The question behind it is simple: for every marketing decision, what proportion requires human judgement versus AI execution?

That ratio must be deliberately set. It should not be left to default. It should not be decided by whichever platform was bought last quarter. It should not be delegated to the loudest person in the room, saying, “Can’t AI do this?”

Different parts of marketing need different levels of human involvement.

Some areas require a high human-to-agent ratio. These include brand positioning, creative strategy, crisis communications, cultural campaigns, ethics calls, product narratives, executive thought leadership, and brand purpose work. AI can assist in these areas through research, synthesis, scenario planning, and option generation. But it should not lead the work.

These are judgement-heavy spaces. The cost of getting them wrong is not just poor performance. The cost can be reputational damage, cultural misreading, loss of Trust, or long-term brand dilution.

Then some areas need a moderate human-to-agent ratio. These include content strategy, audience segmentation, campaign briefs, message testing, performance reviews, competitive analysis, customer journey mapping, and social listening synthesis.

Here, AI can be a very strong assistant. It can help teams move faster, identify patterns, create first drafts, pressure-test assumptions, and organise complex information. But humans still need to decide what matters, what is on-brand, what is commercially useful, and what deserves action.

There are also areas where a lower human-to-agent ratio makes sense. Content drafting, A/B testing, scheduling, SEO support, reporting summaries, media variations, CRM workflow support, and email subject line testing are good examples. In these places, AI can reduce manual burden and help teams become more efficient.

But even here, I would be careful about removing human review entirely. The lower the human involvement, the stronger your governance, brand guidelines, quality controls, and escalation rules need to be.

Finally, there are areas where AI can lead more safely when guided by clear guardrails. Bid management, dynamic creative optimisation, personalisation at scale, campaign pacing, performance signal detection, martech plumbing, and certain data workflows are examples.

This is where marketers should be spending more of their AI energy. Not because these areas are glamorous, but because they are already built around data, optimisation logic, automation, and feedback loops.

Performance marketing, programmatic systems, segmentation, personalisation, and dynamic optimisation already had machine logic before generative AI became the boardroom topic of the year. Generative AI can improve these workflows, especially when paired with clean data, clear governance, and measurable business outcomes.

That is practical adoption.

Slide From My Keynote at the Malaysian Marketing Conference 2026, Read post here

The Visibility Paradox: Backend AI Is Often Accepted Faster Than Frontend AI

One pattern I am seeing quite clearly is this: the more backend the AI use case, the more acceptable it tends to be. The more consumer-facing and emotionally expressive the AI output, the more scrutiny it invites.

That should not surprise us.

Consumers may be comfortable with AI helping to optimise product recommendations, improve customer service routing, personalise the website experience, summarise information, or streamline transactions. These are largely utility-led experiences.

But when AI starts pretending to express empathy, creativity, humour, cultural sensitivity, or brand purpose, the audience applies a different standard.

That is why I worry when brands rush into fully AI-generated video, synthetic people, artificial influencers, or emotionally loaded storytelling without a strong strategic reason.

Yes, AI can support production. It can help with storyboards, visualisation, editing, localisation, versioning, background generation, mock-ups, and ideation. These are useful applications. In many cases, they can save time and cost.

But replacing the human element entirely is a different decision.

A brand should ask: Does AI make this content more useful, more trusted, more distinctive, or more effective?

If the only answer is that it makes it cheaper and faster, that is not a marketing strategy. That reads more like a procurement argument than a marketing strategy.

The Broader Consumer Context Matters

We also need to recognise that consumers are not evaluating AI content in isolation.

They are seeing layoffs. They are hearing leaders talk about replacing jobs. They are watching creative industries struggle with ownership, consent, originality, and fair compensation. They are scrolling through feeds filled with synthetic content, some of it harmless, some of it deceptive, and much of it forgettable.

So when a brand publishes low-effort AI content, the reaction may extend beyond that single asset. It may also be about what the asset represents.

To some audiences, it signals that the brand took a shortcut. To others, it signals that the brand no longer values human creativity. To some, it simply feels emotionally empty.

This is especially relevant with younger audiences, who are often highly fluent in digital culture and quick to detect when content feels manufactured. They may not always object to AI itself, but they object to being patronised by content that lacks authenticity, care, or cultural understanding.

That does not mean brands should avoid AI.

It means they should avoid careless AI.

There is a difference.

The Other Side: AI Discovery May Be A Better Investment Than AI Content Volume

Here is the paradox.

Consumers may be sceptical of AI-generated brand content, but they are increasingly using AI systems for discovery.

They are asking AI tools for product comparisons, recommendations, summaries, buying guidance, category explanations, alternatives, and decision support. This is already changing how people discover brands, evaluate choices, and form consideration sets.

That creates a different opportunity for marketers.

Instead of only asking, “How do we generate more AI content?”, we should also be asking, “How do we make our best content more discoverable, understandable, and trustworthy in AI-mediated environments?”

That requires a different mindset.

It means investing in clear website content, strong product explanations, credible FAQs, case studies, customer reviews, third-party testimonials, expert perspectives, structured information, natural-language answers, and a consistent brand narrative across owned properties.

This is where areas like AEO and GEO become relevant. But again, I would be careful with the mindset. The objective should not be to manipulate AI answers. The objective should be to make your brand easier to understand, verify, and cite.

In simple terms, be more useful to humans and more legible to machines.

That is a better strategy than flooding the internet with average AI-generated content and hoping something gets picked up.

Measurement: Volume Is Not The Win

One of the most dangerous AI metrics in marketing is volume.

Number of posts generated. Number of variants created. Number of campaigns launched. Number of assets produced. Number of tokens consumed. Number of prompts written.

These are activity metrics.

They are not business outcomes.

In the AI era, marketers need to become much sharper about measurement because machines can now create almost infinite activity. The question is not whether something was produced. The question is whether it created value.

Did trust increase? Did brand distinctiveness improve? Did product understanding increase? Did customer acquisition cost improve? Did conversion quality improve? Did community advocacy increase? Did churn reduce? Did incrementality improve? Did the campaign move a commercial or reputational outcome?

If a team produces five times more content but brand trust declines, that is not productivity. That is a synthetic scale with negative leverage.

I have written before that marketing measurement must move from visibility to consequence. AI makes that shift urgent.

When activity becomes cheap, judgement becomes more valuable.

What I Would Tell A CMO Today

If I were advising a CMO on where to start with AI in marketing, I would not begin with “let us automate all content.”

I would start with the workflows where AI can improve performance without weakening trust.

First, look at areas where automation already exists. Performance marketing, dynamic content optimisation, programmatic bidding, CRM workflows, segmentation, analytics, reporting, and campaign operations are sensible starting points. These areas already run on data, rules, optimisation logic, and feedback loops. AI can make them better.

Second, build the governance layer before scaling content. Before producing AI-assisted brand content at scale, lock down your brand voice, claims policy, legal review, compliance process, human approval workflow, disclosure principles, image and likeness rules, cultural review, crisis escalation paths, and data usage boundaries.

Do this before the machine starts publishing, not after.

Third, apply the HAR Framework. Decide where your marketing function needs high human judgement and where AI can take on more execution. Make the ratio explicit. Do not let every team invent its own rules.

Fourth, keep humans in the loop for all brand-facing content. Any content that affects trust, positioning, reputation, culture, product meaning, or brand affinity needs human review. I would make that a default rule.

Fifth, invest in AI discovery, not just AI production. Make your own content clearer, more credible, more structured, and more useful. AI systems can only discover and summarise what they can understand.

Finally, measure outcomes, not output. Do not celebrate content volume. Celebrate trust, conversion quality, brand lift, pipeline influence, customer retention, and community strength.

That is where marketing earns its seat.

In Summary, here is my take

I am optimistic about AI in marketing.

Very optimistic.

It will help good marketers become sharper, faster, more analytical, and more adaptive. It will eliminate repetitive work that many teams should no longer do manually. It will improve performance workflows, research, insights, testing, and execution fluency.

But I also believe we are going to see a correction.

The industry will rediscover the value of human creativity. Not because AI failed, but because marketers will realise that AI works best when it augments the craft, not when it replaces the soul of the work.

The future is not human versus AI.

The future is knowing when the human must lead, when the machine should assist, and when the machine can execute safely within guardrails.

That is the marketing leadership challenge now.

Not adopting AI. Anyone can adopt AI.

The real challenge is knowing where to place it, where to limit it, and where to keep the human firmly in charge.

Lastly, I want to be the first to call it. Very soon, we see a reversal in hiring not only in marketing but also in other parts of the business. This will happen when clarity about how AI fits the enterprise and sensible minds prevail.