Building a transparent map of AI ethics to protect artists' work

We're launching Tally, a benchmark that tracks how different AI models handle creative rights and style copying. Why? Because generative AI models are silently shifting their policies on copying artist's styles right under our noses.

While major AI models like ChatGPT refuse to copy famous artists, they often still mimic lesser-known creators' and internet artists' styles. This double standard extends across mediums – they won't copy renowned painters but will freely imitate writing styles of authors.

This inconsistency exists partly because we lack systematic ways to track and prove style infringement, relying mostly on anecdotes and manual checks. As it stands, only major corporations or wealthy individuals with legal resources can effectively protect their creative rights.

The problem is industry-wide. Companies that push ethical boundaries often gain more users, forcing others to follow suit. This creates a race to the bottom that benefits no one.

Our goal isn't to criticize AI companies – in fact, we believe transparency benefits everyone! When companies have clear data about industry practices, they can make informed decisions, without fear of losing market share to competitors who might be less ethical. A transparent playing field protects both creators and AI companies from the current race to the bottom.

Using Crowdsourcing, we plan to document which artists and styles are being protected versus which are not. Help us by submitting info about artists you believe are vulnerable to being copied by AI, whether they're photographers, writers, artists, or anyone in the creative space at all!