AI in marketing

How Should a University Disclose Its AI Use?

By Midya U · Midya U Advisory

In brief: No federal law in Canada requires a university to disclose AI use in its marketing. Ad Standards' influencer guidance does in some cases, and the EU AI Act's Article 50 transparency rules have applied since 2 August 2026. Disclosing will not make you look worse; denial and copy that reads machine-written are what cost you. Publish a plain-language statement beside your privacy policy that says where you use AI, where you do not, who reviews it, how you label it and when you last checked it.

In the summer of 2026 a track called "Rubberz" reached the Billboard Hot 100. Fenix Flexin, the rapper credited on it, spent weeks denying that AI was involved, before eventually acknowledging it had been used in making the song. Due to the song's success, naturally it generated a lot of controversy.

Your marketing sits in the same gap. What do we say about our own AI use, and when do we disclose it? Nobody is going to hand you the rule book. Write it yourself while writing it is still voluntary.

Do I have to disclose AI-generated content in Canada?

No federal law requires it. The industry code does, in some cases.

No AI law replaced the Artificial Intelligence and Data Act after it died in January 2025, and the privacy bill tabled in June 2026, C-36, says nothing about labelling marketing content. What applies is self-regulation: Ad Standards' October 2025 influencer guidance, which covers AI-generated images and AI influencers. I went through what it means for your photography and your student ambassadors in Should Your Viewbook Ever Use an AI-Generated Photo?

Europe drew a harder line. The EU AI Act's transparency obligations under Article 50 have applied since 2 August 2026. Deepfakes have to be disclosed, and AI-generated text published to inform the public on matters of public interest has to be labelled unless a person has reviewed it and someone holds editorial responsibility for it. Penalties reach €15 million or 3% of worldwide turnover.

What counts as AI-generated content?

The range runs from spell check, through a first draft a person rewrote, a generated image of a residence room, to a student who does not exist. Nearly everyone agrees the last two need a label and the first does not; the disagreements are all in the middle.

The test I use: does the content make a factual claim about your institution that an applicant would rely on? If it does, it has to be real. Everything else can be generated, as long as you say so.

Will disclosing AI use make us look worse?

No. What cost Fenix Flexin was the denial. The track had already charted.

Your audience is further along than the question assumes. HEPI's 2026 student survey, run by Savanta in December 2025 across just over a thousand full-time UK undergraduates, found 95% using AI in at least one way, up from 66% two years earlier.

What does cost you is copy that reads like a machine wrote it. EAB surveyed more than 5,000 US high school students in late 2025, and more than half said they would react negatively to messages they believed were AI-generated. Capture Higher Ed's 2026 survey of about 2,800 US juniors and seniors put it at 60%. Both asked about messages that felt generated, neither tested a labelled one, and both are American. My read is that the risk sits in the giveaway voice, which is what a brand guardrail is for.

Applicants and parents already assume some of what you publish has machine help. The seventeen-year-old reading your program page has almost certainly asked ChatGPT about you before they got there.

I will go first. I use AI throughout my own work, including the first pass of writing this blog post, and I would rather say that plainly than let someone infer it. My own statement is at midyau.com/ai-disclosure, and it covers the same five things below.

What should an AI disclosure statement say?

Put it where your privacy policy sits. Same shelf, same footer link, same stable URL that does not move when the site gets rebuilt. The difference is that almost nobody reads a privacy policy, and this one you want read, so write it the way you would explain it to a parent on the phone.

Five things:

  1. Where you use it. Name the categories: draft copy, translation, image generation, ad variants, the chatbot on your admissions page.
  2. Where you do not. Student testimonials, photographs of real people and places, and any statement about cost, admission requirements or graduate outcomes.
  3. Who reviews it. One named role that signs off before anything goes out. A role rather than a person, so it survives someone taking another job.
  4. How you label it. The exact wording you use and where it sits on the page or the post.
  5. When you last checked it. A date at the bottom of the page and the role that owns it, the way a policy page carries a "last updated" line. Then put a reminder in someone's calendar to actually check and update as needed.

You can call it a statement. If you call it a policy, it acquires a committee, a legal review and a governance calendar, and you will get it back sometime after the next admissions cycle. A statement describing what your own marketing team does is a description of your practice. Publish it, tell your general counsel it is there, and revise it when they come back with notes.

How do I find out where AI is already in our marketing?

Ideally, you should already know the answer to this question. But if you do not, take everything you published in the last 90 days - ad creative, social posts, program pages, recruitment emails, whatever your chatbot has been telling people - and ask whoever made it whether a model touched it, and how.

I'd check two places first, because nobody thinks of them as content: the chatbot on your admissions page, which writes new answers every day, and any page that was translated, because machine translation is AI and is rarely labelled.

Finding where AI has already entered the work, then deciding what you tell people about it, is where a Marketing Function Audit & Reset starts.

Finding where AI has already entered your marketing is the first step of a Marketing Function Audit & Reset.

Book a discovery call →