AI Voices in Podcasts: What Apple Podcasts and Spotify Require

Both allow AI voices. Apple Podcasts now requires you to say so, prominently, in the episode itself and in the show and episode descriptions. Spotify has no disclosure rule, but removes podcasts that copy another creator's voice without permission.

Rules checked on 24 October 2026, from Apple's and Spotify's own guidelines.

What does Apple Podcasts require?

Apple's content guidelines (rule 1.11) say: "Creators using AI to generate audio or video content — including synthetic voices, AI-generated hosts or on-screen personas, or AI-synthesized replicas of real individuals — must prominently disclose this to audiences in the content and metadata for each episode and show."

That's stricter than it was. Until 2026 the rule applied only when AI made up "a material portion" of the audio; now any synthetic voice counts. In practice:

• Say it in the audio. A line in the intro does it: "Our narration is produced with AI voices."
• Say it in the show description and in each episode's description.
• Don't mislead. Rule 1.12 bars using AI "to mislead or deceptively portray real-life events".

What does Spotify require?

Spotify doesn't ask you to label AI voices. Its line is impersonation: "Spotify will remove podcast content that impersonates another creator or host's likeness without their permission — whether that's using AI cloning or any other method". Its platform rules also ban synthetic media presented "as authentic in ways that pose the risk of harm".

So an AI narrator, or a clone of your own voice, is fine on Spotify; a clone of another podcaster isn't. Since you're publishing to Apple as well in most cases, write one disclosure and use it everywhere.

Can I use a clone of my own voice for my podcast?

Yes, on both. On Apple it still counts as a synthetic voice, so disclose it. A clone of someone else's voice needs their clear permission; Leelo asks you to confirm consent every time you clone a voice.

How loud should an AI-voiced podcast be?

Apple asks for "around -16 dB LKFS, with a +/- 1 dB tolerance", with true peaks no higher than −1 dB. Spotify documents only the level it plays music at (−14 LUFS) and normalizes what it plays, so −16 LUFS works on both. Leelo exports at the podcast loudness you choose, and you can measure any file with the free loudness checker.

What kinds of AI podcast work?

• Narrated shows: stories, history, explainers, written by you and read by one consistent voice.
• Two-host conversations written from your notes or a report: see turning a document into a two-host podcast.
• Intros, outros and ad reads in a fixed voice, while you host the rest yourself.
• Translated editions of an existing show, dubbed in your own voice.

What listeners and platforms push back on is the same thing: mass-produced shows with nothing to say, and voices that pretend to be someone real.

All platforms at once: can you use an AI voice on YouTube, Udemy, Audible, Spotify, Steam and TikTok?