How much do AI-generated ad images cost brand trust?

By Q3 2026, the budget argument around ai generated images for ad campaigns has moved past production speed. The harder question is whether cheaper image output stays cheap after consumers notice it. The cleanest warning comes from IAB and Sonata Insights: 82% of ad executives think consumers feel positive about AI-generated ads, while only 45% of consumers actually report positive sentiment. That is a 37-point gap, wider than the 32-point gap reported in 2024.[1]

That survey is useful as a planning signal, not a final verdict. Its sample was 505 consumers and 104 ad executives, so it should not be treated like a nationally representative read on every category. Still, the direction is hard to ignore because the same study also found that 71% of Gen Z and Millennial consumers have seen an AI ad, up from 54% in 2024.[1] This is no longer a niche creative operations decision hidden inside the studio; it is visible to the audience.

A broader trust dataset points the same way. Edelman’s 2026 AI special report, based on 31,000 respondents across 28 countries, found that 61% of consumers say they can identify AI-generated ad content. Among those who identify it, brand trust scores drop 22 points on a 100-point scale.[2] The important word there is “identify.” The penalty is attached to perception, not merely to whether a model touched the asset somewhere in the workflow.

The Trust Penalty Is Now Revenue-Adjacent

For media buyers, trust damage is not a soft concern when it lands on premium perception and purchase intent. Digital Applied’s 2026 benchmark reported that when users perceive an ad as AI-generated, premium perception drops 17%, inspiration drops 19%, and purchase intent drops 14%.[3] That is the kind of number that belongs in the same room as CPM savings, creative throughput, and landing-page conversion rate.

The Digital Applied benchmark is self-published and not peer-reviewed, even though it describes a large base of more than 50,000 ad variations.[3] That caveat matters. A vendor or platform benchmark should not be treated as clean independent evidence. But the finding is still directionally useful because it translates consumer suspicion into metrics a budget owner recognizes: less premium feel, less inspiration, less stated intent to buy.

Klaviyo and Datalily’s consumer survey, covered by EMARKETER in December 2025, adds a simpler sentiment check: only 7% of 8,000 consumers said AI-generated marketing makes them trust a brand more, while 31% said it makes them trust a brand less.[4] That does not prove a specific sales loss. It does make it difficult to defend the assumption that AI-generated creative is brand-neutral by default.

The gap is sharper for younger audiences. IAB reported that 39% of Gen Z consumers have negative sentiment toward AI ads, compared with 20% of Millennials.[1] That should not turn into a lazy generational morality play. It is a targeting warning. If a campaign is built around younger consumers, especially in a category where identity, taste, or status does the selling, the creative savings need to be weighed against a higher probability of resistance.

The CFO Question Is Not “Can We Make It Cheaper?”

The easy spreadsheet line is production cost. AI-generated variants can reduce concepting time, expand testing volume, and keep a stale creative calendar from waiting on another shoot or design queue. Those are real advantages. The incomplete version of the argument stops there and treats every saved production dollar as recovered margin.

A better budget defense asks what happens downstream. If an AI-generated image lowers premium perception, a luxury brand may need more impressions to produce the same qualified demand. If it lowers inspiration, a category that depends on aspiration may see weaker saves, shares, or assisted conversion. If it lowers purchase intent, the media team may end up defending a cheaper asset that made the traffic less valuable.

This is where CTR-only case studies become dangerous. A curiosity click on an uncanny image can look like efficiency in the ad platform while still creating lower-quality consideration. The studies available in 2026 do not let anyone calculate a universal “AI image tax” across every vertical, but they do justify one operating assumption: fully automated AI image production should not be treated as automatically brand-safe simply because it performs acceptably on a surface engagement metric.

Disclosure Helps, But It Is Not a Clean Escape Hatch

Disclosure is the most tempting mitigation because it gives the team something concrete to do. IAB found that 73% of consumers said knowing an ad was AI-created would either increase or not change their purchase likelihood.[1] That is useful in stakeholder conversations because it suggests consumers are not automatically rejecting every disclosed AI-assisted ad.

But the caveat has to be said out loud: this is self-reported purchase likelihood, not observed behavior. Consumers can endorse transparency in a survey and still react differently in a feed, on a product page, or during a high-consideration purchase. Disclosure should be planned as a trust practice, not sold internally as a guaranteed performance fix.

The placement and wording also matter, and the available data does not give execution-level rules. A small “AI-assisted” note in a disclosure area is not the same consumer experience as making generative AI the whole campaign concept. A transparent process page for a technical audience is not the same as a paid social image for a prestige product. The evidence supports disclosure planning; it does not support pretending one label solves every perception problem.

Hybrid Creative Is the Safer Default

The strongest counterweight to a simple anti-AI reading is Deloitte Digital’s 2026 Creative Effectiveness Index. Across 1,840 campaigns, Deloitte reported that hybrid human-AI campaigns outperformed fully automated AI campaigns by 41.3% in brand equity and outperformed fully human campaigns by 29.7% in short-term conversion.[5]

That finding should be handled with the same discipline as any consultancy benchmark. Deloitte has an obvious interest in workflows that require strategic design, human review, and implementation support. Still, the result matches the practical reality many creative and media teams already see: AI is often most useful when it accelerates ideation, resizing, versioning, background exploration, and testable variation, while humans keep control over taste, claims, brand codes, and the final decision to ship.

The budget implication is straightforward. If the choice is between a fully manual workflow that cannot produce enough variants and a fully automated workflow that may carry a measurable trust penalty, hybrid production is the defensible middle. It gives finance some efficiency gain without asking brand, legal, or media teams to absorb the full risk of machine-made creative being recognized as such.

Where the Data Still Does Not Go Far Enough

The available studies are strong enough to reject the idea that AI-generated ad images have no brand cost. They are not strong enough to price that cost for every account. The missing pieces are the ones a media buyer would want before changing a full creative policy: vertical breakouts, price-point sensitivity, observed conversion behavior, repeat exposure effects, and clearer differences between obviously synthetic imagery and AI-assisted images that are hard to detect.

A high-consideration B2B campaign, a premium beauty launch, a discount retail promotion, and a mobile game install campaign should not be expected to carry the same trust risk. The studies available in 2026 do not give clean enough cuts by category or execution style to make that claim. Until they do, teams should avoid turning broad consumer sentiment data into a universal ban or a universal green light.

The practical move is to treat AI-generated imagery as a variable that needs measurement, not as a production shortcut that sits outside performance accountability. Premium perception, brand trust, lead quality, assisted conversion, and post-click behavior belong in the readout when AI-generated assets are tested. If the only reported result is a CTR lift, the test is under-instrumented.

A Decision Frame for 2026 Campaigns

For low-stakes concepting and internal exploration, AI image generation is already hard to argue against. It can widen the creative option set before a team commits production money. For live paid media, especially premium or high-consideration campaigns, the standard should be higher: human creative oversight, intentional disclosure decisions, and enough measurement to catch damage that will not show up in production invoices.

The clearest 2026 read is not that AI-generated ad images are unusable. It is that they are not free in the way the production line makes them look. The trust gap between advertisers and consumers is widening, the penalty touches premium perception and purchase intent, and hybrid workflows appear to capture much of the efficiency without treating brand equity as collateral damage.

When stakeholders ask for AI-generated ad images, the media buyer’s answer should not start with fear or enthusiasm. It should start with the budget math: do the savings survive a measurable hit to premium perception, purchase intent, and trust, and can a hybrid workflow deliver most of the efficiency without absorbing the full brand-equity penalty?

References

1. The AI Ad Gap Widens, IAB
2. Edelman Trust Barometer AI Special Report 2026, Edelman, 2026
3. AI Ad Creative Benchmark 2026 CTR ROAS Data, Digital Applied, 2026
4. Shoppers aren’t impressed by AI-generated marketing, EMARKETER, December 2025
5. Deloitte Digital Creative Effectiveness Index 2026, Deloitte, 2026