AI Persuasion Faces a Transparency Test as 66% Would Feel Manipulated in a Consumer Survey

A successful purchase can still leave a consumer uneasy when the role of artificial intelligence comes to light afterward. A survey of 446 U.S. adults points to disclosure timing as a meaningful part of how people judge an advertiser.

The survey presents a practical problem for brands using AI in advertising: consumers may like the product and still reduce their trust in the company when they learn that AI created the advertisement after the purchase. Respondents who were told about AI before buying reported stronger trust and repeat purchase intent than respondents who learned about it later.

The results also show that disclosure alone will not settle every concern. Across the two equally sized groups, roughly 66% said they would feel manipulated by the advertisement at least somewhat. About 90% said brands should disclose when an advertisement was created entirely with AI, and 88% said a person should review and approve that advertisement before publication.

The 66% figure comes from the survey’s manipulation question. Respondents were divided into two groups of 223 people. One group saw a scenario with upfront disclosure. The other learned after the purchase that AI had created the advertisement and that the ad had not disclosed that fact.

Across both groups, 62.3% of those told upfront said they felt manipulated a lot or some. The figure rose to 70.4% among those who learned about the AI use afterward. Because each group had the same number of respondents, the combined result works out to about 66%.

That pooled number describes a broad concern. The group difference adds a more specific finding: later discovery was associated with an 8.1 percentage point increase in people who felt manipulated. That difference was directionally important, although it did not remain statistically significant after the analysis adjusted for all eight survey comparisons.

The strongest adjusted differences involved three other measures:

The survey therefore separates two related ideas. Many consumers already feel some discomfort about persuasive advertising made entirely with AI. That discomfort becomes more pronounced on several commercial measures when the disclosure arrives after the decision.

A survey of 446 U.S. adults found that consumers reacted more favorably when an advertisement’s use of AI was disclosed before purchase rather than revealed afterward. Upfront disclosure was associated with higher trust in the advertiser and stronger repeat purchase intent, even though many respondents still felt manipulated or misled. The results suggest that brands should treat AI disclosure as part of the consumer decision—not a post-purchase explanation—and pair it with meaningful human review. Because the survey used hypothetical scenarios and separate nonprobability samples, the findings offer a practical warning rather than definitive proof that disclosure timing alone caused the differences.

The survey used the same basic purchase story for both groups. Each respondent imagined seeing an online advertisement for a product they had not planned to buy. The ad persuaded them to make the purchase, and they were satisfied after using the product.

The timing of the AI information changed the scenario:

That last detail matters. The later scenario included a clear nondisclosure statement, so respondents were reacting to both the timing of the information and the fact that the brand had left it out. The design reflects a realistic consumer complaint, but it also means timing is not the only factor shaping the result.

The product outcome adds another useful detail. Every respondent was told they were satisfied with the purchase. A positive product experience therefore existed in both scenarios, yet it did not produce equal levels of trust or loyalty.

Product satisfaction may protect a brand from some complaints. In this survey, it did not erase concern about how the advertising was made or when that information became available.

The clearest result appeared in the difference between immediate trust and future behavior.

Among people who saw the disclosure before buying, 76.7% said they would trust the advertiser a great deal or a fair amount. Among people who learned about the AI use afterward, 65.0% gave one of those favorable responses. The gap was 11.7 percentage points, with a 95% confidence interval ranging from 3.3 to 20.0 points in the direction of lower trust after later discovery.

Repeat purchase intent followed the same pattern. Nearly four in five respondents in the upfront group, or 79.8%, said they would be very or somewhat likely to buy from the advertiser again. The later-discovery group reached 67.3%, a 12.6 point difference.

That result has a direct business implication. A campaign can meet its immediate objective by persuading a consumer to buy. If the consumer later feels that important information was withheld, the same campaign may weaken the next purchase opportunity.

The recommendation measure moved in the same direction, although the difference was smaller and less precise. About 73.1% of the upfront group said they would be very or somewhat likely to recommend the advertiser, compared with 67.7% of the later-discovery group. The adjusted result did not meet the survey’s threshold for statistical significance.

The distinction matters for marketers evaluating campaign performance. Clicks, conversions and immediate sales capture the first response. Trust, repeat purchase and recommendations capture what happens after the creative process becomes part of the consumer’s judgment.

The survey’s most useful lesson is not that every AI-generated advertisement will damage trust. It is that hiding or delaying material information can make the creative process part of the consumer’s judgment. A visible, specific disclosure gives people context before they act, while meaningful human review signals that the brand remains accountable for the message. Marketers should therefore evaluate AI campaigns on more than conversion: trust, repeat purchase intent, complaints and post-purchase reactions reveal whether short-term persuasion is creating a longer-term credibility cost.

Upfront disclosure produced more favorable results, but it did not create universal comfort. In that group, 57.0% still said they felt misled a lot or some, and 62.3% said they felt manipulated at least somewhat.

Several parts of the scenario could explain that response. The purchase was unplanned, and the advertisement persuaded the respondent to act. Some people may have reacted to the persuasive force of the ad itself. Others may have objected to the idea of an entirely AI-created message or image. The survey did not ask respondents to identify the specific source of their discomfort.

The wording also described a stronger form of AI involvement than many real campaigns use. The scenario said that AI created all of the advertisement’s words and images. A brand using AI to resize an image, create several copy variations or assist a human writer may produce a different reaction.

That distinction should shape how the findings are used. A disclosure policy designed for full AI generation may need different language from one covering editing assistance or automated targeting. Consumers may care about the part of the process that affects the message, the image, the recommendation or the claim they are being asked to believe.

The survey found nearly equal support for AI disclosure and human review. Across all 446 respondents, 90.1% said it was very or somewhat important for a brand to tell consumers when an advertisement had been created entirely with AI. Another 88.3% said it was important to know that a person had reviewed and approved the advertisement before publication.

The human-review result was almost identical across the two groups: 87.9% among people told upfront and 88.8% among people who learned later. That consistency suggests that review is a separate expectation rather than a reaction limited to people who felt deceived.

A useful review process needs a named person or team with enough authority to reject or change the material. A ceremonial click through an approval workflow gives consumers little reason to trust the claim. A meaningful review should cover at least:

The reviewer also needs a record of what the system produced, what was changed and who approved the final version. That record helps the brand investigate a complaint and gives internal teams a way to distinguish AI output from human judgment.

The survey fits with a broader shift toward clearer AI labeling in advertising and marketing.

The Interactive Advertising Bureau’s AI Transparency and Disclosure Framework V2 uses a risk-based and materiality-driven approach. That framing gives brands room to consider how much AI shaped the consumer-facing material, how visible the use is and whether the information could affect a consumer’s interpretation.

The European Commission’s guidance on transparency obligations under Article 50 of the AI Act also places attention on the organization operating the AI system in relevant situations. For advertising companies and brands working across borders, the identity of the responsible organization matters alongside the wording of the disclosure.

Consumer exposure is increasing as well. A 2026 Bentley University-Gallup survey cited in the background research found that 79% of Americans had seen advertisements in the previous 30 days that appeared to be made using AI. A Canva study reported that 70% of consumers believe it will eventually become impossible to tell whether an ad was AI-generated without disclosure.

Those findings create a practical problem for visual judgment. Consumers may learn to expect disclosure even when they cannot identify AI use on their own. A label can therefore become part of the basic information needed to evaluate a message, rather than a special notice reserved for unusual campaigns.

The survey points toward several concrete operating rules for advertisers.

A company should record whether AI generated the words, images, voice, video, targeting recommendation or only a minor production element. The disclosure should match the actual use. Calling a campaign entirely AI-generated when a human wrote the copy creates its own credibility problem. Calling a campaign human-created when AI produced the central message creates a different one.

A disclosure shown after checkout, in a follow-up email or in a later brand story arrives after the consumer has acted. The survey suggests that the timing changes how people evaluate the advertiser.

The notice should appear close enough to the advertisement that a consumer can see it while considering the claim. A vague statement buried in a general privacy policy gives consumers little practical choice. Plain language such as Created with AI and reviewed by [Brand] may be more useful when it accurately describes the process.

A disclosure should be visible on the screen where the advertisement appears, including mobile placements. It should avoid technical language that forces the audience to understand model architecture before learning the basic fact that AI shaped the material.

The exact wording will depend on the campaign. An advertisement with AI-generated imagery may need a different notice from one with an AI-written script and a human actor. The central question is whether the consumer can understand the role AI played before responding to the message.

A reviewer should evaluate the final advertisement as a consumer might see it. That includes combinations of text and image that may look acceptable when checked separately. The reviewer should also have a route for correcting the source model, prompt, asset library or targeting rule when a problem appears.

A brand can show disclosed and undisclosed versions to a research panel, then ask about trust, manipulation, fairness and willingness to buy again. This type of testing will not predict every marketplace response, but it can reveal when a creative concept depends on consumers failing to notice how it was made.

Campaign reporting should include more than conversion. Brands can monitor repeat purchase, customer support contacts, refund explanations, social comments and trust measures after disclosure. A short-term sales lift paired with a drop in confidence deserves attention before the campaign is expanded.

ContentEngine conducted two online surveys through SurveyMonkey Audience. Each tranche included 223 completed responses, for a combined total of 446. One tranche was fielded on August 25 and 26, 2026, and the other on August 25, 2026.

The samples were separate nonprobability samples. Respondents opted into online panels, and the research did not randomly assign people to one scenario within a single experiment. The samples were similar on the measured demographic variables, but those checks do not account for unmeasured differences between the groups.

The results support an association between later discovery and lower trust, lower repeat purchase intent and stronger feelings of being misled. A larger randomized experiment would provide stronger evidence about the independent effect of disclosure timing. That study could hold the wording constant, vary only when the disclosure appears and test several levels of AI involvement.

The scenario was hypothetical. Respondents did not view a real advertisement, spend money or use an actual product. Stated trust and purchase intent can differ from behavior in a marketplace. The survey also did not measure race, ethnicity or education, and it did not ask respondents to explain their answers.

The statistical analysis adjusted the eight comparisons using the Holm method. Trust, repeat purchase intent and feeling misled remained statistically significant after that adjustment. Regret, recommendation, feeling manipulated and the importance measures showed smaller or less precise differences between the groups.

Those limits narrow the claim in a useful way. The findings describe how these survey respondents reacted to a particular scenario involving full AI generation and explicit nondisclosure. They provide a warning for brands, alongside a clear reason for more detailed research on partial AI use, different disclosure language and actual purchasing behavior.

For brands, disclosure is tied to the moment when a consumer decides whether to believe, click, buy or keep listening. That timing gives the information practical value. It lets the audience judge the message with knowledge of how it was produced.

The survey also places human accountability beside transparency. Consumers want to know that AI was involved, and they want evidence that a person remained responsible for the final advertisement. A label without review can leave questions about accuracy and judgment. Review without disclosure can leave consumers feeling that the brand controlled information they deserved before acting.

A persuasive ad therefore carries more than its product claim. It also communicates how openly the brand is willing to explain the tools behind that claim. The purchase in this survey went well, yet later discovery still changed the relationship. For advertisers using AI at scale, that makes the disclosure decision part of the customer experience itself.