Working Professionally
If you are using these tools for money — client work, products, publication, licensing — the questions stop being about prompting and start being about ownership, permission and disclosure. This section is the part most courses skip and most working artists actually need.
This is a practical orientation, not legal advice. The law here is genuinely unsettled and varies by country. For anything with real money attached, ask a lawyer in your jurisdiction.
The short answer in the United States, and broadly similar in the UK, EU and elsewhere: a work produced by AI *on its own* cannot be copyrighted, because copyright requires a human author. A work in which you made substantial creative contributions can be — and it is protected to the extent of what you contributed.
Where this stands in 2026. The Supreme Court declined to review Thaler v. Perlmutter in March 2026, leaving in place the appellate ruling affirming the Copyright Office's human-authorship requirement. That question is settled: AI cannot be an author.
What remains open — and is being worked out case by case — is exactly how much human involvement is enough.
The useful way to think about it. A prompt alone is generally treated as an idea, and ideas aren't protected. But when you use these tools the way you'd use a camera or a brush — making sustained creative choices about the result — you are contributing protectable expression. Practically, that means:
- Typing a prompt and downloading the first result → weakest position, likely unprotectable
- Extensive iteration, selection, masked editing and composition → considerably stronger
- Your own sketch, photograph or linework as the structural basis → stronger again; the underlying drawing is yours outright
- AI elements composited into a larger work you built → your arrangement and your original elements are protected
Keep records — this is the concrete, actionable part. If you ever need to demonstrate authorship, contemporaneous evidence of your process is what does it. Keep:
- Prompts and how they evolved, and why
- Outputs you rejected, and what was wrong with them
- Your own source material — sketches, photos, references
- Screenshots or layered files showing masking, editing and compositing
- Notes on the decisions you made at each stage
This is the prompt journal from the iteration section, doing a second job. It costs almost nothing to maintain and is impossible to reconstruct afterwards.
Copyright is about whether anyone else can copy your work. Licensing is about whether the platform permits you to sell it at all. These are separate questions and people conflate them constantly.
Read your tool's terms for these four things specifically:
- Are commercial rights included on your plan? Several platforms grant them only on paid tiers — free-tier output can be personal use only.
- What happens if you stop paying? Some terms retroactively restrict use of work made during a subscription once it lapses.
- Are your generations public? Several platforms make everything visible to other users by default, including on paid plans. This matters enormously for client work under NDA — check before you generate anything confidential.
- Is there indemnification? Some enterprise offerings (Adobe, Google, OpenAI, Microsoft) will defend you if a third party claims your generated output infringes. For meaningful commercial exposure, this is worth real money.
Client contracts. Increasingly, agency and publisher contracts contain explicit AI clauses — sometimes disclosure requirements, sometimes outright prohibition, sometimes a warranty that you own everything you deliver, which you may not be able to give truthfully for purely generated work. Read them, and raise it early rather than at delivery.
Stock and platform rules differ too. Some stock libraries accept AI work with disclosure, some ban it, some require model releases you cannot obtain for a generated face. Print-on-demand services, app stores, game marketplaces and competitions all have their own policies. Check the destination before you make the work.
Most large image models were trained on web-scraped material, including a great deal of work by living artists who neither consented nor were paid. This is the field's central ethical problem, it is the subject of most of the 70-plus active lawsuits, and if you make your living from art it is not somebody else's issue.
Where the litigation actually sits: essentially all of it targets model developers, not the people using the tools. As a working artist you are not the defendant — but the outcomes will shape what tools exist and on what terms, and at least one significant appellate ruling is expected soon.
What you can actually control:
- Prefer models with cleaner provenance where the work warrants it. Adobe Firefly (licensed and public-domain training data, with indemnification) and ElevenLabs Music (licensed catalogue from the start) are the clearest current examples.
- Don't prompt living artists by name. It is unreliable, several platforms restrict it, and it is the specific practice that damaged the relationship between this technology and the art community. Describe the mechanics of a style instead — as covered earlier, it works better anyway.
- Train on your own work, not other people's. Custom style training is offered by several platforms; feeding it your own portfolio produces a model that is yours in every sense.
- Be honest about what you did. Passing off generated work as hand-made damages you far more than disclosure ever will, and it is increasingly detectable.
On the anxiety. These tools are very good at producing competent images and consistently poor at knowing which image is worth making. Taste, judgement, art direction, knowing what a client actually needs, and knowing when something is finished — none of that is automated, and all of it is what you were being paid for. The artists doing well with these tools are using them to move faster through the parts that were never the interesting bit.
The difference between playing and working is repeatability: being able to quote a job, hit a deadline, and deliver something consistent. A workflow that holds up under a client brief looks roughly like this.
- Establish the look before you produce volume. Small, cheap explorations until you have three or four images that define the direction. Get sign-off on those.
- Lock your references. Style reference, character reference, palette. Everything afterwards is generated against them.
- Bring your own structure in. Sketch the compositions, or shoot them, or block them out. Use guided generation or image-to-image so the layouts are yours. This is what makes the set feel authored rather than assembled — and it is what strengthens your authorship claim.
- Generate in batches, then curate hard. Expect to keep a small fraction. The curation is the art direction.
- Fix rather than reroll. Inpaint, edit conversationally, composite. Rerolling a nearly-right image is how deadlines get missed.
- Upscale and finish in your normal tools. Grade, retouch, typeset, deliver.
- Archive the process — prompts, seeds, references, versions — with the job files. You will need it for the revision round, for the next job in the series, and potentially for an authorship question later.
On pricing. Price the outcome, not the hours, and be careful about competing on speed alone — that is a race where the tools get faster every quarter and you don't. What clients are paying for is judgement, consistency, revisions and someone accountable for the result. None of that got cheaper.