Skip to content
AI & Credentials

AI Content Policy

What AI may do in a course and what it may not, the declaration every course makes before publishing, synthetic instructors, and what ACADA itself uses AI for.

Last updated on October 25, 2026.

Use AI to help you make a course; do not let it make the course. AI that drafts, edits, translates or captions needs no mention. Material AI made that you publish as the teaching, and any AI voice, avatar or persona that teaches, must be declared in Course Settings, and the landing page says so before anyone enrolls; a course cannot be published until you have answered. A course that is mostly machine-generated is not your own teaching and is not allowed, declared or not. Whatever AI wrote, you published: you check it, you source it, you fix it. Nothing anyone posts on ACADA trains an AI, and no decision about your account is made by one.

This summary is for orientation only. The sections below are the operative terms.

1. Who this applies to

This policy applies to everyone who publishes on acada.ac, run by ACADA L.L.C., a Tennessee limited liability company (ACADA, we), and mostly to instructors and collaborators, because it is about courses. It says what AI may do in a course, what must be disclosed and where, and what we use AI for ourselves. How students may use AI in assignments is in section 4 of Academic Integrity; who owns AI-made material is in section 3 of Intellectual Property; what the ACADA Assistant is, as advice, is in the Educational Disclaimer. Everything in Prohibited Content applies however content was made.

2. What we expect

A course on ACADA is taught by a person. AI can help that person plan, draft, edit, illustrate and check; it cannot replace them. What you publish is your statement to your students, and it carries your name whatever tool produced the words. When AI made something a student will learn from, the student is told.

3. Three kinds of AI use

  • AI-assisted. Drafting or rewriting text you then work over, editing, translating, captioning, transcribing, help with code, touching up an image, generating ideas or an outline. No disclosure. It is a tool, like a spellchecker.
  • AI-generated material. Lesson text, images, diagrams, audio, video or quiz questions that AI produced and you published as they are, as part of the teaching. Disclosed on the course under section 4.
  • A synthetic instructor. An AI voice, avatar or persona that presents some or all of the teaching, including a clone of your own voice or face. Disclosed on the course under section 4 and never presented as a real person, under section 6.

The test is what the student receives: if a person read, checked and stands behind every word, it is assisted; if the machine's output is the lesson, it is generated.

4. Declaring it

The rule. Course Settings asks whether the course contains AI-generated material or a synthetic instructor, with a note of what AI made. You must answer before the course can be published; the site refuses to publish a course that has not declared. The landing page shows the declaration before anyone enrolls, and nothing is shown for a course with no AI-generated material. A course published before this policy shows nothing until its instructor answers, which is required the next time it is published. Keep the declaration current: when you add AI-made material, raise it; when you replace it with your own, you may lower it. Where we find AI-generated material that was not declared, we place the disclosure ourselves, marked as ours, and you can raise it or edit the note but not lower it.

Example of a violation. A course whose diagrams and narration are AI-made, declared as containing none.

Consequence. We place the disclosure and tell you; it is recorded in your Account standing, where you can appeal it. Leaving it off again, or declaring none when the material plainly is AI-made, brings a written warning and the ladder in section 9.

5. The teaching must be yours

The rule. Section 5 of the Course Content Policy says a course must be substantially your own teaching. AI-generated material may be a part of a course, declared; it may not be the course. A course whose lessons are mostly what a model produced from your prompts, however good, is not your teaching and is not allowed, with or without the disclosure.

Example of a violation. Forty lessons generated from a topic list and published with a synthetic narrator, the instructor having written the prompts and nothing else.

Consequence. The course is unlisted until it is rebuilt around your own teaching, and removed if there is nothing of yours to rebuild around; a written warning, and the ladder in section 9 after that.

6. Synthetic instructors and real people

The rule. A clone of your own voice or face may teach, declared under section 4. A fictional AI persona may teach, declared, and never presented as a real person: no invented credentials, no biography that reads as a life lived. An AI likeness of any other real person, their voice or their face, needs that person's consent, and presenting one as the real person is impersonation under the Community Guidelines and Account & Identity. The name on your account stays yours under Account & Identity; a persona is a character in the course, not the instructor.

Example of a violation. A course "taught by Dr. Elena Marsh, 20 years in oncology", where Dr. Marsh is a generated face reading generated text.

Consequence. The persona's claims are removed and the disclosure placed; a written warning. Passing a synthetic instructor off as a real person to sell a course is fraud under Fraud & Scams and closes the account at once.

7. Accuracy, sources and grading

The rule. AI output you publish is your statement. You read it, check it, and cite where its facts come from as Academic Integrity requires, and you fix a confirmed error within the time that policy sets. "The AI said so" is not a defence. No AI grades anything on ACADA: a quiz is scored against the answers and the pass mark you set, an assignment counts as done once it is submitted, and a certificate follows when every lesson is complete. If you use an AI tool to help you give feedback on submissions, the feedback is yours: you read the submission, and you stand behind what you say.

Example of a violation. A lesson with a fabricated statistic and a citation to a paper that does not exist, left in place after a student's correction was confirmed.

Consequence. Handled under Academic Integrity, section 7: the correction, then the ladder.

8. Images, thumbnails and everything outside a course

The rule. AI-made thumbnails, covers, diagrams and illustrations need no disclosure, with two limits: no realistic AI image presented as a real event, place or person, and no likeness of a real person without their consent. Outside courses: a review written by AI is not the student's own account and is treated as fake under Reviews & Ratings; a press article substantially written by AI says so at the top; messages, bios and room posts have no rule beyond the Community Guidelines. The ACADA Assistant may help you plan and improve a course, and anything it drafts that you publish is AI-generated material under section 3.

Example of a violation. A thumbnail that looks like a photograph of the instructor with a famous chef, generated; a five-star review the student had a model write.

Consequence. The image or review is removed and the ladder in section 9 follows.

9. How we enforce this

  • The ladder. For anything in this policy we follow the Account Enforcement & Appeals policy: a written warning, then restrictions, then closure of the role or the account. A missing disclosure is placed first; a course that is not your teaching is unlisted until it is, or removed.
  • At once. A synthetic instructor passed off as a real person to sell a course is fraud and closes the account the day we confirm it.
  • Appeals. Every decision under this policy can be appealed under that policy.
  • Reporting. Report a course with the reason "Spam or misleading" and say what is undisclosed, or write to support@acada.ac. A person reviews every report within 3 business days.

10. What ACADA does with AI

The ACADA Assistant is the only AI that ACADA runs. It answers questions about the platform and about the course you have open, it runs on Anthropic's API, and it explains rather than writing work a student will hand in. Choosing a language from the language menu uses Google's own translator, as the Cookie Policy says; it translates the page you are reading and is not ours. Nothing anyone posts on ACADA, courses, messages, submissions, reviews or notes, is used to train AI models, ours or anyone else's, as the Intellectual Property policy sets out. No decision about your account or your content is made by an AI: a person reviews every report, every copyright notice and every appeal, and the Assistant has no part in it.

11. Changes to this policy

When we change this policy, we change the date shown at the top. If a change adds a rule or a consequence, we post a notice in the app at least 14 days before it takes effect.

12. Related policies