Lesson objective
Convert the FTC staff description of an endorser training and monitoring programme, PRSA's recommended practices and the IAB summary into target behaviours, labelling what each source binds and whom.
Plain-language explanation
Why this matters now
You do not have to invent the behaviours you teach. Several bodies have already described what a reasonable control looks like in marketing and communications work. Those descriptions are ready-made target behaviours. They also come with a discipline you must keep: each source binds different people in different ways, and you have to say which.
This lesson works through three: the FTC staff description of an endorser training and monitoring programme, PRSA's recommended practices for members on AI agents, and the IAB's AI Transparency and Disclosure Framework V2, read only as its summary page. Each becomes a small set of behaviours. None becomes a legal safe harbour.
Professional guidance (FTC staff, not binding): "FTC's Endorsement Guides: What People Are Asking". The FAQ says the Endorsement Guides do not have the force of law, but practices inconsistent with them can result in enforcement under section 5 of the FTC Act. It describes a reasonable training and monitoring programme with four elements: instructions to endorsers, guidance on disclosures, periodic searches of what network members are saying, and corrective action. It treats simple disclosures such as "#ad" and "Ad:" as workable, and "#endorsement" or "#ambassador" alone as inadequate. The page shows no date in the fetched content, and it does not address AI-generated or virtual endorsers.
Professional guidance (PRSA, advisory for members, not law): ESA #22, "AI Agents in Public Relations", July 2026. Issued by PRSA's Board of Ethics and Professional Standards, it states that "AI systems are tools, not ethical actors." Under the Code's Disclosure of Information provision, "AI agents must not conceal their identity, sponsorship, or intent." Its recommended practices are to be transparent, maintain human oversight, monitor for bias, avoid deceptive practices, build AI literacy and governance (including agency-client policies), and assess security risks. It lists improper practices including undisclosed AI identity, hidden influence and manipulation, synthetic front groups, and lack of human accountability.
Industry guidance (IAB, voluntary): AI Transparency and Disclosure Framework V2, dated 18 August 2026. The summary page says V2 translates the original framework's risk-based, materiality-driven approach "into practical protocols for when and how AI involvement should be disclosed" across consumer-facing advertising and marketing content. It says V2 covers AI-generated and AI-assisted text, imagery, video, audio, synthetic voices, digital twins and AI-powered consumer interactions, offers common disclosure principles and use-case triggers, and aims to balance transparency "with avoiding unnecessary labeling and disclosure fatigue." The V2 document itself was not read for this programme. You may teach only what the summary page says.
A question to sit with
A learner asks: "If we follow the IAB framework and the FTC's checklist, are we safe?" What is the honest answer, and how do you keep it from sounding like a shrug?
What goes wrong
1. Staff guidance is taught as law. "The FTC requires a monitoring programme" overstates a staff FAQ. Say instead that FTC staff describe what a reasonable programme looks like, that the Guides do not have the force of law, and that practices inconsistent with them can lead to section 5 enforcement.
2. A professional code is taught as binding everyone. PRSA's advisory guidance speaks to PRSA members. It is useful to anyone, but it binds no non-member.
3. An industry framework is taught as a safe harbour. The IAB framework is voluntary industry guidance. Following it does not create a legal safe harbour.
4. The summary is taught as the document. An instructor lists "the IAB V2 triggers" from memory or from a colleague's slide. This programme read only the summary, which says triggers exist and does not list them. Teaching a trigger list you have not read is inventing it.
5. The source is stretched past its subject. The FTC FAQ does not address AI-generated or virtual endorsers. A slide that says "the FTC FAQ says AI influencers must disclose X" is attributing to the page something it does not say.
The practical answer
Best practice: turn each element of a control description into one observable behaviour. The FTC staff description converts almost directly.
| FAQ element | Target behaviour for a creator manager | Evidence in the session |
|---|---|---|
| Instructions to endorsers | Write a creator brief that states what the creator may and may not claim | The brief's claims section |
| Guidance on disclosures | Write the disclosure into the brief, using a clear form such as "#ad" or "Ad:", never "#ambassador" alone | The brief's disclosure section |
| Periodic searches of what network members say | Run and log a search of a fictional creator's recent posts | The search log with dates |
| Corrective action | Write the correction request for a fictional post that fell short | The written request |
Best practice: give PR rooms behaviours from the PRSA recommended practices, labelled for members.
| Recommended practice | Target behaviour | Label |
|---|---|---|
| Be transparent | Write the disclosure of AI involvement into an AI-assisted outreach plan | PRSA advisory guidance for members |
| Maintain human oversight | Name the practitioner who evaluates each AI output before it is sent | PRSA advisory guidance for members |
| Avoid deceptive practices | Check a fictional outreach plan for undisclosed AI identity or a synthetic front group, and strike it | PRSA advisory guidance for members |
| Build AI literacy and governance | Draft the AI clause for a fictional agency-client policy | PRSA advisory guidance for members |
Best practice: teach IAB V2 as a place to look, not a list to recite. Because only the summary was read, the target behaviour is procedural: "Before deciding whether AI-assisted content carries a label, check the current IAB V2 document and the client's own policy, record which you checked and the date, and route any disagreement to the account lead." That behaviour is honest about what you know. It also reflects the summary's stated aim of balancing transparency with avoiding disclosure fatigue: the learner decides with the document open, not from memory.
Best practice: never let a source's silence become a rule. Where a source does not address something, as the FTC FAQ does not address virtual endorsers, the behaviour is to note the gap and route the question to the account lead and counsel. Teach the gap as part of the content.
AI Coalition Network recommendation: put the binding label on every behaviour card. Each behaviour card in your session carries a line saying where it comes from and what that source is: staff guidance, a member code's advisory guidance, a voluntary framework, or this programme's standard. Learners should never have to guess whether they are hearing law.
Remember this
- Real control descriptions are ready-made target behaviours; convert each element into something a learner does.
- The FTC FAQ is staff guidance: not law, but practices inconsistent with the Guides can lead to section 5 enforcement.
- PRSA's ESA #22 is advisory guidance for PRSA members, not law for everyone.
- IAB V2 is voluntary with no safe harbour, and only its summary was read: teach it as a place to look, never as a trigger list.
- Label every behaviour card with what its source is and whom it binds.
Marketing agency example
Aisling Moreau, a fictional instructor, is hired by the fictional Brightwater Social agency to train its six creator managers and three PR staff. The agency's old slide says: "FTC rules: creators must use #ambassador and the IAB V2 triggers." Aisling checks each part. The FTC staff FAQ treats "#ambassador" alone as inadequate, and it is staff guidance, not a rule. The IAB V2 triggers are not described on the summary page this programme read, so she cannot teach them. She deletes the slide. For the creator managers she builds four behaviour cards from the FAQ's description of a training and monitoring programme: write the claims section of a brief, write the disclosure section using a clear form such as "#ad", log a search of a fictional creator's posts, and write a correction request. Each card carries the label "FTC staff guidance, not binding; departures from the Guides can lead to section 5 enforcement." For the PR staff, two of whom are PRSA members, she builds three cards from ESA #22: name the practitioner who evaluates each AI output, write the AI disclosure into an outreach plan, and strike an undisclosed AI persona from a fictional plan. Each card is labelled "PRSA advisory guidance for members." Six creator managers with four cards each and three PR staff with three cards each means twenty-four plus nine, thirty-three completed cards from those two sources to review. For the IAB framework she adds one card for all nine learners: check the current V2 document and the client's policy before deciding on an AI label, and record which you checked and when. A creator manager asks whether a fictional virtual influencer must disclose that it is AI. Aisling says the FTC FAQ does not address virtual endorsers, and routes the question to the account lead and counsel.
Responsible-use warning
Never teach the FTC staff FAQ as a rule, PRSA's advisory guidance as binding non-members, or the IAB framework as a legal safe harbour. Never list IAB V2 triggers you have not read; only the summary page was read for this programme. Never attribute to the FTC FAQ anything about AI-generated or virtual endorsers, which it does not address. Never present "#ambassador" alone as an adequate disclosure.
Human-review requirement
The instructor owns every behaviour card and its source label. Before delivery, a second reviewer checks each card against its source, confirms the binding label is accurate (staff guidance, member advisory guidance, voluntary framework or this programme's standard), confirms no IAB trigger is described beyond the summary page, and confirms nothing is attributed to a source that does not say it. Counsel owns questions the sources leave open. The agency's named reviewers own every real creator brief and outreach plan.
Practical exercise
Convert control descriptions into labelled behaviour cards. Use a fictional agency, fictional creators and fictional briefs throughout.
- List the four elements of the endorser training and monitoring programme described in the FTC staff FAQ.
- For each element, write one target behaviour for a creator manager, with an observable verb and the evidence it leaves.
- Choose three of PRSA's recommended practices from ESA #22 and write one target behaviour for a PR practitioner for each.
- Write one procedural behaviour for the IAB V2 framework that does not describe any trigger beyond the summary page.
- Put a binding label on every card: staff guidance, member advisory guidance, voluntary framework or AI Coalition Network recommendation.
- Find one question these sources leave open, such as disclosure by a virtual influencer, and write the referral a learner should make.
- Build one fictional creator brief with a planted disclosure error, such as "#ambassador" alone, and check that one of your cards would catch it.
- Count your cards and confirm the total matches the number of elements and practices you converted.
- Write the answer you would give a learner who asks, "If we follow the IAB framework and the FTC checklist, are we safe?", keeping each source's binding label accurate and ending with the referral for anything the sources leave open.
Knowledge check
-
What four elements does the FTC staff FAQ describe for a reasonable endorser training and monitoring programme?
Answer guidance: Instructions to endorsers, guidance on disclosures, periodic searches of what network members say, and corrective action.
-
How should PRSA's ESA #22 be labelled when used in a session?
Answer guidance: As advisory guidance from PRSA's Board of Ethics and Professional Standards for PRSA members, not as law binding everyone.
-
Why may an instructor not list the IAB V2 use-case triggers?
Answer guidance: Because only the summary page was read, which says triggers exist but does not list them; teaching a list would be inventing it.
-
Does following the IAB framework give an agency a legal safe harbour?
Answer guidance: No. It is voluntary industry guidance and does not create a legal safe harbour for disclosure decisions.
Sources and further reading
- Federal Trade Commission — FTC's Endorsement Guides: What People Are Asking. Accessed 2026-09-15.
- Public Relations Society of America (PRSA), Board of Ethics and Professional Standards — AI Agents in Public Relations, Ethical Standard Advisory (ESA) #22 (2026-07-01). Accessed 2026-09-15.
- Interactive Advertising Bureau (IAB) — AI Transparency & Disclosure Framework V2 (web page titled 'AI Transparency & Disclosure Standards V2') (2026-08-18). Accessed 2026-09-15.
- AI Coalition Network — Certification Educational Standard: Human Review of AI-Assisted Professional Work. Accessed 2026-09-15.