A language model can write an artist bio in ten seconds. The problem isn’t the speed: it’s that, without direction, it produces exactly the kind of text a festival programmer discards at the first sentence — full of adjectives that say nothing, structured like the thousand other bios the same model has already read.

Used well, an AI assistant (ChatGPT, Claude, Gemini) is a useful tool for speeding up your first draft and generating variants. Used poorly, it produces generic text that needs to be rewritten from scratch. The difference lies in how you use it, not in whether you use it.

Where AI genuinely helps

Going from long to short (and back)

If you already have a 250-word bio you're happy with, ask it to condense it to 70 words while keeping the facts, not the filler phrases. This works far better than asking it to write from scratch.

Generating variants for different contexts

One version for the press, one for social media, one for a grant application. Give it your base bio and ask for tone adaptations, not new content.

Spotting weak sentences

Paste your bio and ask explicitly: "which sentences in this text don't provide concrete information?" It's a better detector of empty adjectives than you are, because it has no emotional attachment to the text.

Fixing third-person consistency

If you've written a draft in first person and need to convert it to third (as a professional bio should be), AI does it accurately and in seconds.

Where it fails, and why that’s dangerous

It invents credentials. If you give it too little information, a language model tends to fill in the gaps with plausible but false details: festivals you never played, training you never did. Never accept a sentence containing a concrete fact (date, venue, institution name) without verifying it yourself.

It loves the empty adjective. “A multifaceted artist with a unique voice that challenges genre boundaries” is exactly the kind of sentence a model generates by default when you don’t give it enough concrete material. It’s the same language any programmer discards in three seconds — and models produce it because they’ve learned from thousands of equally generic bios.

It doesn’t know your target audience. A model has no idea whether the programmer reading your bio is looking for “renewal of traditional repertoire” or “electronic innovation.” You need to give it that context explicitly, or the result will be generic by definition.

The practical rule

Use AI for structure and speed, never for facts. Every concrete piece of information — names, dates, venues, figures — must come from you and be verified before you publish. If the model proposed it and you don't recognise it, cut it.

A workflow that actually works

  1. Write the basic facts yourself first, without worrying about style: name, discipline, two or three real credentials, what sets you apart.
  2. Feed it to the AI with a concrete instruction: “structure this as a professional third-person bio, 70 words, no generic adjectives.”
  3. Ask it to question the result: “review this text and flag any sentence that doesn’t provide verifiable information.”
  4. Read it yourself, out loud if needed, and cut any sentence that could belong to any other artist.

The result isn’t a bio written by AI. It’s a bio written by you — faster to produce and cleaner of empty adjectives thanks to an assisted editing process.

For the remaining criteria — what the first sentence must contain, the difference between a short and long bio, what should never appear in the text — see the full guide: The artist bio: how to write one that works.

If you’d like us to review your bio and the rest of your press materials, tell us about your project.