IL-2023-000007 - [2025] EWHC 2863 (Ch)

It is explained that the Model is “intended for research purposes only” and possible research areas and tasks are then identified, namely (a) safe deployment of models which have the potential to generate harmful content; (b) probing and understanding the limitations and biases of generative models; (c) generation of artworks and use in design and other artistic processes; (d) applications in educational or creative tools; (e) research on generative models.

Under the heading “Misuse and Malicious Use” it is explained that “[t]he model should not be used to intentionally create or disseminate images that create hostile or alienating environments for people. This includes generating images that people would foreseeably find disturbing, distressing or offensive; or content that propagates historical or current stereotypes”. Misuse is said to include (but is not limited to): (a) generating demeaning, dehumanizing, or otherwise harmful representations of people or their environments, cultures, religions, etc; (b) intentionally promoting or propagating discriminatory content or harmful stereotypes; (c) impersonating individuals without their consent; (d) sexual content without consent of the people who might see it (e) mis- and disinformation; (f) sharing copyrighted or licensed material in violation of its terms of use; and (g) sharing content that is an alteration of copyrighted material or licensed material in violation of its terms of use.

Under the heading “Out of Scope Use” it is expressly stated that “[t]he model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out of scope for the abilities of this model”.

Under the heading “Limitations” it is stated (amongst other things) that “the Model does not achieve perfect photorealism” and that “the model was trained on a subset of a large scale dataset LAION-5B which contains adult, violent and sexual content. To partially mitigate this, we have filtered the dataset using LAION’s NSFW detector”.