Gains:
- Ability to understand the prohibition of misleading advertising, Advertising Board regulations and honesty/evidence principles and audit the artificial intelligence output for legal compliance.
- Brand safety, transparency of artificial intelligence-generated content, and the ability to understand ethical boundaries and eliminate risky claims and images in advance
- Being able to protect that the legal and ethical responsibility of advertising cannot be transferred to artificial intelligence and that the final approval always belongs to the advertiser.
Advertising is a creative profession as well as a responsible one. An advertisement affects the purchasing decision of millions of people; This power brings with it the obligation of honesty. In the age of artificial intelligence, this responsibility increases, not decreases: AI can produce striking but unsubstantiated claims, realistic but misleading images, and outrageous promises in minutes. This final unit deals with the legal and ethical framework of advertising. The basic principle is clear from the beginning: The accuracy, honesty and legality of advertising is the responsibility of the advertiser; This responsibility cannot be delegated to artificial intelligence. Saying "AI suggested" will protect you neither in an audit nor in a court of law.
Misleading advertising ban and Advertising Board
Commercial advertisements in Türkiye are regulated by the Consumer Protection Law No. 6502 and the related Commercial Advertising and Unfair Commercial Practices Regulation. The board that implements and supervises these regulations is the Advertising Board. The Advertising Board may impose sanctions such as suspension, correction and administrative fines for misleading or deceptive advertisements. On the broadcast (TV/radio) side, RTÜK regulations also include; KVKK comes into play in the use of personal data.
The main prohibition is misleading advertising: advertising that deceives or may deceive the consumer as to the attributes, price, results or source of the product. Misleading does not have to be intentional; As a result, if it misleads the consumer, it is prohibited. Two related principles are honesty (the advertisement must be truthful) and burden of proof (the advertiser is obliged to prove every objective claim made in the advertisement). So, if you say "it repairs the skin by 90 percent", you have to prove it with valid and up-to-date evidence.
The following table shows common risky claims and their safe alternatives:
Risky phrase
Why is it risky?
safe alternative
"The best in Türkiye"
unprovable superiority
"Our customers love" (based on)
"Exact result, 100%"
Absolute/unproved promise
"Supports with regular use"
"Prevents diseases"
Health claim without evidence
Approved, compliant expression
"number one"
Flimsy sorting
A concrete, sourced data
"Free" (with hidden condition)
Hidden conditional spoofing
State the condition clearly
Beware: AI is prone to recommending statements of superiority and certainty without evidence just because they have high sales power. Treat every word “most,” “precise,” “100 percent,” “prevents,” “guaranteed” in the printout as a red flag and subject it to an evidence/regulatory check.
Ethics: honesty, transparency and care
The law is the minimum line; Ethics is broader than that. A few basics of advertising ethics: Honesty — toeing the line between exaggeration and lies. Transparency — not disguising that a piece of content is advertising (covert advertising, unlabeled sponsorship is problematic); It is expected that content produced with AI will increasingly be mentioned. Care for vulnerable groups — not exploiting vulnerable groups such as children and those with health problems. Avoiding discrimination — Not reproducing stereotypes about gender, ethnicity, etc. in texts and images produced by AI. AI can reflect biases in the training data; It is the advertiser's job to monitor the output from this perspective.
Brand safety
Brand safety is to ensure that the advertisement does not appear next to content media that will harm the brand's reputation. In automatic ad purchasing (programmatic) systems, your ad may appear next to a page containing hate speech, violence or disinformation; This wears out the brand. This risk grows as automation with AI increases. Brand safety; It is protected by exclusion lists, appropriate content categories, and regular moderation. A related concept is brand suitability: choosing to appear in brand-appropriate contexts, not just avoiding harm.
Step by step: pre-publication legal-ethical review
Step 1 — Claim inventory. List each objective claim in the ad.
Step 2 — Proof check. Is there valid, up-to-date evidence for each claim?
Step 3 — Precision/superiority screening. Mark expressions such as "most," "certain," "100 percent," "prevents" and make them defensible.
Step 4 — Transparency and precision. Is it obvious that it is advertising, should AI production be specified, is there exploitation of vulnerable groups and stereotyping?
Step 5 — Brand safety. Where will the advertisement be published and from which media should it be excluded?
three mini cases
Case 1 — Proof saves. A cosmetics brand wanted to use the phrase "eliminates wrinkles" suggested by AI. Upon legal review, this was found to be unprovable and misleading. The team translated it to "helps reduce the appearance of wrinkles" based on product testing. The ad went on air safely. The proof check prevented an Advertising Board sanction.
Case 2 — Risk of covert advertising. One brand worked with a creator but did not label the post as “advertisement/collaboration.” A complaint was filed alleging covert advertising. Lesson: it should be transparently stated that a piece of content is an advertisement; Hiding it is both unethical and subject to sanctions.
Case 3 — Trademark safety violation. A brand's programmatic ad appeared on a site that spread disinformation through automated placement. When the screenshot spread on social media, the brand suffered a loss of reputation. Lesson: automation doesn't eliminate brand safety oversight; exclusion lists and regular checking are essential.
Four copyable templates
1) Claim and proof checker:
Examine the ad copy below: "[paste text]".Task: (1) List each objective claim (requiring evidence) in the text.(2) Write what type of evidence would be required for each claim.(3) Mark those that are difficult/impossible to prove and suggest a conservative alternative.Not legal advice; Generate a list for me to check.
2) Red flag scanner:
Text: "[ad copy]".Task: Flag phrases that pose a risk of misleading advertising: superiority without proof ("best", "number one"), absolute promise ("sure", "100%", "guarantee"), health/financial claim without proof, "free/free" with hidden condition. For each mark, suggest a defensible alternative.
3) Transparency and precision control:
Content: "[text/image description]". Place of publication: [channel]. Task: Check: (1) Is it obvious that this content is an advertisement, is a label required? (2) Do I need to indicate that it is content produced with AI? (3) Is there an aspect that exploits sensitive groups such as children/patients? (4) Does it create stereotypes on issues such as gender/ethnicity? Write what I should do for each item.
4) Brand safety checklist:
Campaign: [description]. Publication: programmatic/social/[...]. Task: Create a checklist for brand safety: categories of content to exclude, appropriate/inappropriate contexts, frequency of placement moderation, and steps to follow in case of complaint. Add a heading of "environments to avoid" based on the brand's values.
Weak prompt / Strong prompt
Weak prompt:
Write a best-selling ad copy that praises my product in the strongest possible way.
This demand invites unsubstantiated superiority and exaggerated promises; The risk of misleading advertising is high.
Powerful prompt:
Product: skin moisturizer. Publication: Turkey, digital.Task: Write a persuasive but LEGAL and HONEST ad copy.Rules: use superiority without evidence ("best") and absolute promise ("sure result");make health/treatment claims; Make each benefit statement moderate and defensible. Next to it, list what evidence I will need for each claim in the text.
The second request observes both persuasive and legal limits and requires a match between claim and evidence.
Common mistakes
- Delegating responsibility just because "AI suggested it". Legal and ethical responsibility always lies with the advertiser and the brand.
- Using superlative/certainty without evidence. “Best,” “100 percent,” “prevents” create the burden of proof; Without evidence is misleading.
- Hide the ad. Covert advertising and unlabeled sponsorship are subject to sanctions.
- Ignoring vulnerable groups and stereotypes. AI can produce bias; The output should be inspected from this perspective.
- Leaving brand security to automation. If programmatic placement goes unchecked, the brand appears in risky environments.
- Mistaking health/financial statements for ordinary claims. These areas are subject to special and strict rules.
Tip: Before publishing an ad, ask yourself: "Can I document this claim in an audit? Is it clear that this content is an ad and reflects fact?" If you cannot answer "yes" to both questions, the text is not ready yet.
In summary
In addition to creativity, advertising carries the obligation of honesty and legality, and in the age of artificial intelligence, this responsibility increases. Commercial advertisements in Türkiye are inspected by the Advertising Board within the framework of Law No. 6502 and the Commercial Advertisement Regulation; Misleading advertising is prohibited and every objective claim creates the burden of proof. Ethics is broader than law: transparency, care for vulnerable groups, avoidance of stereotypes and brand safety are part of this framework. AI can produce striking but risky expressions; It is the advertiser's job to pass each output through the filter of proof, accuracy/superiority, transparency and brand safety. Responsibility can never be transferred to artificial intelligence.
Application task
Choose an ad copy (either your own product or an AI-generated draft). (1) With the “Claim and proof checker”, list each objective claim and required evidence in the text. (2) Use the “red flag scanner” to flag superlatives/certainties and turn them into defensible alternatives. (3) Audit ad tag for AI specification and stereotype with “Transparency and sensitivity check”. (4) Create a "brand safety checklist" for where the campaign will run. (5) Finally, answer the question "can I defend this claim with documentation" for each claim.
checklist
- [ ] I listed every objective claim in the ad and checked the proof.
- [ ] I replaced unproven superiority and absolute promises with measured statements.
- [ ] I have transparently stated that the content is advertising; I marked the AI production if necessary.
- [ ] I oversaw care for vulnerable groups and risk of stereotyping.
- [ ] I prepared an exclusion and audit plan for brand safety.
- [ ] I internalized that the responsibility belongs to me and the brand, not to the AI.
Module Exam
1. An advertising team publishes a campaign text produced by artificial intelligence directly without any checks. What is the fundamental mistake in this approach?
- A) Converting the artificial intelligence output into publication without subjecting it to brand, originality and legal control; Ignoring that the responsibility lies with people ✔
- B) It is strictly forbidden to use artificial intelligence in advertising
- C) Artificial intelligence always writes the text too long
- D) The text is not presented in a table
Description: Artificial intelligence produces drafts and suggestions; However, the originality of the text to be published, its suitability to the brand voice, the evidence of the claims and its legal compliance are the responsibility of the advertiser. An unverified printout is like an unapproved campaign and carries the risk of misleading advertising and copyright.
2. Which statement best explains why the creative brief is critical in idea generation with artificial intelligence?
- A) Brief is for presentation purposes only, it does not affect the output quality
- B) Brief gives the artificial intelligence the purpose, audience, message and constraints, ensuring that the output is brand-appropriate and focused ✔
- C) Once the brief is written, the AI output no longer needs verification
- D) Brief completely eliminates the copyright problem of artificial intelligence
Description: The creative brief defines the campaign's purpose, target audience, main message, tone and constraints. The better the context given to artificial intelligence, the more accurate it produces ideas; Without a brief, the output will be generic and disconnected from the brand.
3. What is the most effective way to keep brand voice consistent across AI outputs?
- A) Trying a different tone in each text and choosing the one you like best
- B) Never defining the brand voice and relying on the default tone of artificial intelligence
- C) Turning the brand voice guide (tone, words, don'ts, examples) into a system instruction and using it in every production ✔
- D) Just limit the slogan length
Description: When a system instruction/guide is given to artificial intelligence, where the brand voice is defined by tone, word preferences, prohibited expressions and examples, the outputs are consistent. Writing a prompt from scratch every time produces inconsistency.
4. Which of the following is true for a persona produced with artificial intelligence?
- A) It is an accurate and verified profile of real customers
- B) It is a document that, once produced, should never be updated
- C) Completely eliminates the need for segmentation
- D) It is a draft hypothesis that needs to be tested with real data and field verification ✔
Description: An artificial intelligence persona is a draft of assumptions derived from available data and generalizations. Real customer data cannot be considered accurate without testing it through surveys and field observations; otherwise the wrong message will go to the wrong audience.
5. Which statement most accurately defines the concept of 'value proposition' in advertising?
- A) A statement that clearly explains the concrete benefit the product offers to the target audience and the reason for its preference ✔
- B) Long list of product specifications
- C) Description of the brand's logo and colors
- D) List of channels where the advertisement will be broadcast
Description: The value proposition clearly expresses the concrete benefit the product offers to the target audience and why it should be preferred over its competitors. It is centered on the benefit (what it brings to the customer), not the feature (what the product is).
6. What is the best way to use artificial intelligence when producing advertising text for the same product on different channels (Google search, Instagram, e-mail)?
- A) Producing a single text and copying it exactly to all channels
- B) Giving the format, constraints and user intention of each channel to artificial intelligence and producing and editing adapted variations ✔
- C) Putting the longest text everywhere without considering the channels
- D) Just change the number of emojis according to the channel
Description: Each channel has different character limits, user intent and format. The right approach is to give artificial intelligence channel constraints and intent to produce adapted variations, and then edit them. It is ineffective to replicate a single text everywhere.
7. What is the most critical point of attention in terms of copyright and brand when preparing advertising visuals with generative visual artificial intelligence?
- A) The image should be as high resolution as possible
- B) Always use blue color in the image
- C) Verifying that the output does not violate the rights of another brand/work/person and complies with the brand identity ✔
- D) The image must be produced in one go
Description: Generative images are influenced by training data; Prints that imitate another brand's logo, the face of a recognizable person, or a copyrighted work pose legal risks. It is mandatory to verify the outputs in terms of copyright, personality rights and brand consistency.
8. A media planner is evaluating the budget allocation suggested by artificial intelligence. How should one handle this suggestion?
- A) As a final plan that is precise and ready to be implemented
- B) As an output that has no value and should be ignored completely
- C) Only if the total budget is correct, the details will be displayed as an unimportant table.
- D) As an initial scenario that needs to be verified against actual cost and performance data ✔
Description: The channel mix and budget distribution suggested by artificial intelligence is a starting scenario. Actual platform costs (CPM/CPC) cannot be applied without verification against historical performance and targets; The final decision belongs to the planner.
9. Which of the following is the basic rule in a healthy A/B test?
- A) Isolating a single variable and keeping the others constant and not declaring results before reaching sufficient samples ✔
- B) Changing the title, image and button at the same time
- C) Look at the data in the first few hours and finish the test immediately
- D) Choosing the winner based on personal taste rather than statistics
Explanation: In A/B testing, changing a single variable (for example, just the title) and keeping the others constant helps understand the reason for the difference. Changing many things at once makes it unclear which element is effective. Additionally, adequate sampling and significance are essential.
10. What does the ROAS (Return on Ad Spend) metric measure?
- A) Total number of people who saw the advertisement
- B) Return on advertising expenditure; Ratio of revenue generated to advertising expenditure ✔
- C) How many seconds an advertisement is watched
- D) Total number of pages on the website
Description: ROAS measures return on ad spend: the ratio of revenue generated to ad spend. For example, if 40,000 TL of revenue was generated with 10,000 TL of expenditure, ROAS is 4. It is used to evaluate the profitability aspect of the campaign.
11. In a campaign report, artificial intelligence produced a number saying 'the click-through rate increased by 35 percent compared to last month'. What should the advertiser do?
- A) Since it is written fluently, accept the number as correct and add it to the report.
- B) Round the number and make it appear higher
- C) Not putting the number in the report without comparing it with the source data (platform/analytics dashboard) and verifying it ✔
- D) Putting it in the presentation without indicating the source
Explanation: Language models can produce fluent but made-up numbers (hallucinations) when they cannot access data or calculate it incorrectly. Each number and causality sentence should not be included in the report without being compared and verified with the source data (advertising platform, analytics dashboard).
12. Which board in Türkiye inspects misleading advertising and commercial advertisements and can impose administrative sanctions in case of non-compliance?
- A) Central Bank
- B) Competition Authority Press Council
- C) Capital Markets Board
- D) Advertising Board ✔
Description: The Advertising Board inspects commercial advertisements within the scope of the Consumer Protection Law No. 6502 and the Commercial Advertising and Unfair Commercial Practices Regulation; It may impose sanctions such as suspension and administrative fines for misleading or deceptive advertisements.
13. The phrase 'Türkiye's best cream' is wanted to be used in an advertisement, which cannot be proven. What is the right approach?
- A) Not to use if the claim cannot be supported by independent evidence; ✔ replace with a measured and provable statement
- B) Use it as is because artificial intelligence recommends it
- C) Strengthen the claim further and make it 'the best in the world'
- D) Assuming that there will be no problem if written in small font
Explanation: Unproven claims of superiority (best, number one, definitive result) fall within the scope of misleading advertising and carry the risk of sanctions. Even if artificial intelligence suggests such a statement, if the claim cannot be supported by independent and up-to-date evidence, it should not be used and should be replaced with a measured and provable statement.
14. What does brand safety mean in advertising?
- A) Protecting advertising images with a password
- B) Ensuring that the advertisement is shown in appropriate contexts, away from content environments that will harm the brand ✔
- C) Keeping the advertising budget safe in the bank
- D) Registering the brand logo as copyright
Description: Brand safety refers to ensuring that advertisements do not appear next to content environments that will harm the brand's reputation (hate speech, violence, disinformation, etc.) and that the brand message is placed in appropriate, safe contexts. As automation increases with artificial intelligence, this control becomes even more critical.