How do customers respond when an ad discloses AI use?
What difference does it make to write “Created with AI” in the ad and explain why? We explain the methodology of the three experiments, their different results, and their meaning for the brand team.
ACADEMIC WORK REVIEWED
“Made with AI” but why? How consumers interpret beneficiary-framed AI disclosures in advertising
English review: 9 September 2026 · The source date appears in the citation above.
Editorial review draft
The source has been checked; the review still carries its original editorial status. DijitalPi did not conduct a customer experiment for this paper.
LET'S READ THE RESEARCH TOGETHER
We first explain the researchers' question, method and findings. We then discuss how to interpret the results, clearly separating DijitalPi's commentary from the source.
01 / WHAT DID THE RESEARCHERS WANT TO UNDERSTAND?
Does explaining why AI was used make customers respond more positively?
Schreiner and colleagues investigated whether adding a use justification to the AI description in the ad changed consumer evaluations. The detail of the statement and the emphasis on who it benefits were compared.
Explaining the context · DijitalPi commentary
A brand knows the reasoning behind its content, but the viewer sees only the ad and a few words of explanation. That makes it useful to ask whether the benefit the brand wants to communicate matches what the audience actually understands.
For example, a design team might respond positively to the phrase “To deliver more creative visuals.” A viewer may notice the creative benefit, find the explanation unnecessary or miss it entirely. These are possible reactions that show why an experiment is needed; they are not reported findings from the study.
02 / HOW WAS THE RESEARCH CONDUCTED?
How did they test the question?
Three experiments were conducted in the USA with 339, 359 and 359 participants. In the first, text descriptions were evaluated, and in the other two, Instagram-like ads were evaluated. Advertising and brand attitudes and interaction intention were measured; not a realized sale.
Understanding the method · DijitalPi commentary
Attitude is the evaluation a person gives about an advertisement or brand. Interaction intention is the behavior that the person says he/she intends to do. A person may say they like the ad, but may not need the product. For this reason, we do not derive sales estimates directly from the survey response.
Reading the text alone and seeing it in the advertisement are different conditions. When using a research finding, it is necessary to consider where one's attention is directed. Someone who is specifically asked to review the description and someone who evaluates the visual in the stream are not in the same role.
03 / RESEARCH FINDINGS
The advantage in the text experiment was not replicated in the advertising environment.
In the first experiment, more concrete explanations were evaluated more positively. In two social media-like experiments, annotations did not improve evaluations; made it negative in some comparisons. The emphasis on who the benefit is for was not sufficiently accurate to exclude small effects in these two experiments.
Making a single "it worked" sentence here would lose half the story. It should be explained under what conditions the result was observed. If a title or description is evaluated together with the context in which it is examined, the detail that the research teaches us is preserved.
“No significant difference found” does not always mean “the two options are exactly the same”. A small difference may still be possible if the measurement is not precise enough. In such a case, we maintain the uncertainty stated by the authors and do not produce a winner of our own choice.
Our brand-communication interpretation is that disclosure length alone is not the right measure. A more concrete starting point is whether readers understand the production process correctly. Adding a justification that does not reflect the real process solves nothing; the disclosure must match the record.
05 / CONCLUSION AND OPEN QUESTIONS
What did we learn, and what do we still not know?
The findings do not show that explaining the reason for AI use consistently improves consumer response. The results are limited to a US sample and experimental conditions; they do not measure campaign sales in Turkey.
What does your target audience understand from the expression used? It is necessary to try this separately on the same image and offer, with a predetermined criterion. The following examples illustrate such preparation. The conclusion of the article should not be read as a recommendation to hide the use of AI.
We created these scenarios to make the topic concrete. They are not cases from the paper or measured client results.
EXAMPLE 01
Non-real benefit
The team wants to write “We used AI to protect the environment” in the description. However, there is no study comparing the environmental impact of production.
Would you use this justification because it sounds good?
Open the recommendation for example 01+
Describe only the process that you can verify.
Write clearly at what stage AI is used and who controls the final content. Do not include environmental, economic or social benefits that you have not measured as justification. Make sure the description matches your production record.
EXAMPLE 02
Did everyone understand the label?
A small AI disclosure appears in the corner of the ad. The team thinks it is clear, but viewers were never asked what they understood.
Is finding the label enough to understand?
Open the recommendation for example 02+
Ask the audience to describe it in their own words.
Show the sample image and ask “How was this content prepared?” ask. If it is confusing which part of the image is produced and which is the actual product photo, simplify the explanation. Write down the answers without prompting.
EXAMPLE 03
Likes increased, demand is the same
The new explainer ad got more likes. Despite this, the number of people going to the product page and asking for information did not change.
Would you report this as a sales success?
Open the recommendation for example 03+
Evaluate likes, visits and requests separately.
Determine your goal before you try. When changing the description, don't change the image, audience, and bid at the same time. If there is not enough data, record the observation and uncertainty instead of declaring a winner.
TRY IT WITH YOUR TEAM
Read the description of an ad together.
Have one person from the brand, design and customer communications team review the same ad. Write the role of AI and the promise given to the customer separately. Correct any disagreements before trying a new campaign.
Timo Schreiner, Yakov Bart, Koen Pauwels, Cosima Schütze and Fabian Buder. Nuremberg Institute for Market Decisions; Northeastern University, D'Amore-McKim School of Business.
The academic work belongs to the researchers named above. This page contains DijitalPi's explanatory review and original business examples; it is not a full translation of the paper.
Cookie preferences