How to Create an AI Use Policy for Your Podcast Team
You don't need a legal document. You need a clear set of rules your team can actually follow so your voice, your standards, and your guest relationships stay intact.
"The goal isn't to restrict AI. It's to make sure AI works for your show — not the other way around. A good policy takes you about two hours to write and saves you months of quiet brand damage."
Before you can set rules, you need to know what's actually happening. Most hosts are surprised to find AI already embedded in more places than they realized. Start by asking your team directly and without judgment.
- Send your team a simple form: "List every tool you use that has any AI feature, even if it's just a button you rarely click."
- Map out which tasks each tool is being used for: writing, research, editing, scheduling, social media, or guest outreach.
- Note which of those tasks involve content that goes public under your name or your guest's name.
- Flag anything that touches guest quotes, outreach messages, or episode summaries as high priority.
This is the most important step and the one most hosts skip. Your team cannot protect your voice if they don't know what it is. Write it down plainly, not as a style guide, but as clear rules anyone can apply without having to guess.
- Write three sentences that describe how you sound when you're at your best. Share them with your team as the standard.
- List five phrases or patterns you would never use. These become your team's red flags when reviewing AI output.
- Define what accuracy means for your show. Does a quote need to be verbatim? Does a summary need to capture a specific type of insight?
- Write down which relationships require a personal touch from you, not AI assistance. Guest outreach is almost always on this list.
My show sounds like: [Your 2-3 sentence description. Example: direct but warm, never corporate, always curious. I ask follow-up questions most hosts skip because they're too polite.]
I would never say: [List 4-5 phrases. Example: "game-changer," "dive deep," "at the end of the day," "circle back."]
A quote is only usable if: [Your standard. Example: it matches the transcript word for word, or I have personally approved a light edit.]
Not every task needs the same level of oversight. A traffic light system lets your team move fast on low-risk work while keeping the right things in front of you for review.
- Green (AI can do it, no review needed): internal research drafts, scheduling, transcription, rough first cuts of internal documents.
- Yellow (AI assists, human reviews before use): show notes, episode summaries, promotional social posts, suggested interview questions.
- Red (AI does not touch this): guest outreach emails, published quotes attributed to guests, any content signed with your name as final.
- Put this list somewhere your team sees it daily. A shared doc, a Notion page, or a pinned Slack message all work.
This is your safety net. No matter what category a task falls into, anything that goes public under your name or your guest's name needs a human to sign off. The review doesn't have to be long. It just has to happen.
- Decide who is responsible for final review. It doesn't have to be you for everything, but it does need to be someone with judgment and authority to say no.
- Build the review step into your existing workflow so it's not optional. If you use a project management tool, it becomes a required stage before anything moves to published.
- Create a two-question checklist for reviewers: "Does this sound like us?" and "Would our guest be comfortable with this?"
- Agree that if the answer to either question is uncertain, the content does not go out without further input.
Your policy doesn't need to be long. A single page is enough. What matters is that it exists, your team has seen it, and you treat it as a living document rather than a one-time exercise.
- Write your policy as a one-page document with four sections: what AI is approved to do, what requires human review, what AI must never do, and your voice standard.
- Walk your team through it in a short call rather than just sending it. Answer questions in real time so people actually understand the intent, not just the rules.
- Set a recurring quarterly reminder to review and update. AI tools change fast, and your workflow will evolve.
- When you onboard anyone new to your team, make this document part of day one.
Section 1 — What AI can do without review: [Your green list]
Section 2 — What AI can assist with, human reviews before use: [Your yellow list]
Section 3 — What AI must never do: [Your red list]
Section 4 — Our voice standard: [Your three-sentence description, your banned phrases, your quote standard]
Last reviewed: [Date] — Next review: [Date]