This article is general information for writers, not legal advice. Laws on AI and copyright differ by country and are still developing. For a decision that matters, talk to a qualified lawyer in your jurisdiction.
The short answer
In general, yes. Using ChatGPT or a similar tool to help write a book is not illegal in most places, and OpenAI’s terms, at the time of writing, allow users to use the output, including commercially, subject to its usage policies. But “legal to do” and “free of problems” are different questions. The main issues are copyright ownership, possible infringement, and disclosure rules set by publishers, platforms and industry bodies.
Who owns what ChatGPT writes?
There are two separate questions: what the tool’s terms say, and what copyright law says.
- The terms of service govern your relationship with the provider. They typically address who may use the output and for what purposes. Read the current terms of whichever tool you use, because they change.
- Copyright law decides whether the text can be protected as a work at all, and by whom. A company’s terms cannot create copyright where the law does not recognise it.
Copyright and human authorship
In the United States, the Copyright Office has taken the position that copyright protects works with human authorship. Material generated purely by an AI system, from a prompt, is generally not treated as protected, and the Office has asked applicants to disclose AI-generated content beyond minimal amounts when registering. At the same time, a human can generally claim protection for their own original contributions, such as text they wrote themselves, and for creative selection, arrangement or modification of AI output.
What that means in practice:
- A book you largely write and edit yourself, with AI used for brainstorming or light suggestions, sits on much firmer ground than one made mostly of unedited AI output.
- Relying on long stretches of raw AI text could leave those parts without copyright protection, which matters if you later want to stop others from copying them.
- Other countries take different approaches, and some are still working them out. Do not assume the US position applies to you.
Infringement and training-data questions
Two separate risks are worth understanding:
- Output that resembles existing work. An AI tool can produce text that is similar to material it learned from. If you publish something that substantially copies a protected work, the fact that a tool produced it is unlikely to protect you. Check for passages that look familiar, especially distinctive phrasing, characters or plots.
- Training-data lawsuits. Authors and publishers have brought cases against AI companies over how models were trained. These cases are ongoing or being resolved in different ways, and the law is unsettled. For an individual writer, the practical effect is uncertainty rather than a clear rule.
A related caution: asking a tool to imitate a specific living author’s style or to continue someone’s existing characters and world can raise additional legal and ethical concerns.
Platform and industry rules
Even where something is legal, the places you want to publish or sell may have their own rules:
- Self-publishing platforms. Some, including Amazon KDP, have asked authors to disclose AI-generated content when publishing. Policies change, so check the current rules before you upload.
- Traditional publishers and agents. Many contracts now contain warranties that the work is original and sometimes language about AI. Read yours carefully.
- Contests and magazines. Many have explicit AI policies, and some prohibit AI-generated submissions.
- Screenwriting. The Writers Guild of America’s 2023 agreement set out that AI-generated material is not treated as “literary material” and that a company cannot require a writer to use AI tools. Writers working under such an agreement should check the current terms and talk to the guild if unsure. Producers and buyers of spec scripts may also have their own views on AI-assisted work.
Practical steps for writers
- Do the creative work yourself. Use AI, if at all, for brainstorming, summarising your own notes or catching typos, rather than generating whole chapters or scenes.
- Keep records. Save drafts and notes that show your own contributions and which parts were assisted.
- Check for similarity. Read AI-assisted passages critically and rewrite anything that sounds borrowed.
- Follow disclosure rules. Read the current policy of each platform, publisher or contest before you submit.
- Review the tool’s terms. Look for how the provider handles your input, and avoid pasting in confidential or unpublished material you cannot afford to share.
- Don’t use it to imitate. Avoid prompts that ask for text in the manner of a particular living writer or that continue protected works.
A craft note for screenwriters
Beyond the legal questions, there is a practical one: does the tool help your writing? Characters and dialogue that come from generic patterns tend to sound generic. The qualities that make a script stand out, such as a specific voice, real stakes, and a character with a flaw that costs them, come from decisions only you can make. If you use AI at all, treat it as a sounding board and keep the authorship in your own hands. For the craft side, start with character development for screenwriters.
When to talk to a lawyer
Consider professional advice if you plan to register copyright in a work that involved AI, sign a publishing or production deal with AI-related terms, rely on a tool in a work-for-hire arrangement, or if you receive a claim that your work copies someone else’s. A lawyer who works with intellectual property in your country can give advice that a general article cannot.