Kickstarter mandates generative AI disclosures for creators

Crowdfunding has a trust problem right now, and generative AI just poured gasoline on it. Backers are tired of pledging money to creative projects, only to discover the artwork was spit out by Midjourney in five seconds without anyone disclosing it.

Kickstarter is finally putting its foot down. The crowdfunding giant announced a new policy requiring all projects utilizing generative AI to disclose full details on their campaign pages, set to take effect on August 29.

Show your work and name your data sources

Under the new rules, any creator using AI to generate text, images, audio, or code will have to answer a dedicated questionnaire during submission.

They must outline exactly how AI is incorporated into the workflow, spelling out which elements are entirely original human work and which are machine-generated. This will appear publicly on the campaign page under a new "Use of AI" section.

For projects building generative AI tools or software, the requirements get even stricter. Creators must disclose the sources of their training data, detail how consent and attribution are handled, and demonstrate safeguards such as opt-in or opt-out mechanisms for original artists.

Submissions will go through human moderation, and Kickstarter warned that hiding AI usage could lead to immediate project suspension. Existing campaigns submitted prior to August 29 will not be retroactively penalized.

From cartoon clones to explicit models

Kickstarter has been struggling to balance its creator roots with the AI boom since late last year.

In December, the platform took heat before banning Unstable Diffusion, an unrestricted AI image generator looking to raise funds without safety rails. More recently, Kickstarter mistakenly approved—and then hastily removed—a project that used generative AI to rip off an existing indie comic book.

Susannah Page-Katz, Kickstarter's director of trust and safety, noted that community feedback overwhelmingly demanded transparency. Rather than guessing whether a project is authentic, backers want creators to explain their process in plain English.

Why it matters

For everyday backers, this is a much-needed layer of protection. You can immediately see whether a board game's artwork comes from a freelance illustrator or an algorithm, helping you decide where your dollars go.

For AI tool developers, however, this policy creates a massive hurdle. Even OpenAI refuses to reveal the training data behind its latest models due to legal and competitive paranoia. Expecting grassroots software builders on Kickstarter to provide clean paper trails for massive datasets is ambitious at best.

We think Kickstarter took the most sensible middle ground available. Banning generative AI outright would have been short-sighted and unenforceable; leaving it unregulated was actively poisoning backer trust.

The real challenge lies in enforcement. Automated AI detection is notoriously unreliable, meaning Kickstarter's human review team will bear the brunt of policing ambiguous submissions. If creators choose to simply lie on their disclosure forms, Kickstarter may still find itself playing an endless game of whack-a-mole.