Talk:Artificial intelligence and copyright

Bias?

Doesn't this seem a bit biased: "As of 2023, there were a number of US lawsuits disputing this, arguing that the training of machine learning models infringed the copyright of the authors of works contained in the training data. Commentators have suggested that if the plaintiffs succeed, this may shift the balance of power in favour of large corporations such as Google, Microsoft and Meta which can afford to license large amounts of training data from copyright holders and leverage their own proprietary datasets of user-generated data." ??? It mentions what a commentator speculates might be a (bad for smaller corporations) result if plaintiffs succeed, but doesn't mention what might be a (bad for artists/writers) result if plaintiffs lose. Presumably, although large corporations could better pay for training data, the artists/writers/etc would at least have some hope of getting paid at all whilst having their work ripped off. (And the big corporations are going to hog the field anyway, it's just that they'll have a smaller profit if they have to pay for their training data.) Has no one been a "commentator" on that? 2601:600:9080:D490:A99B:E7DA:A297:1901 (talk) 02:13, 12 June 2023 (UTC)

If you have other sources, feel free to add them to the article, within the usual constraint of WP:Reliable Sources. Rolf H Nelson (talk) 05:02, 21 June 2023 (UTC)

Wiki Education assignment: Technical and Professional Writing

This article was the subject of a Wiki Education Foundation-supported course assignment, between 17 January 2024 and 7 May 2024. Further details are available on the course page. Student editor(s): Virgocat444 (article contribs).

— Assignment last updated by Eaturvegeez (talk) 21:24, 10 February 2024 (UTC)

Potential Edits

Howdy!

I will be editing to eliminate bias and and endure that the facts are correct. I will also be re-structuring the information for readability. Virgocat444 (talk) 17:42, 5 March 2024 (UTC)

Wiki Education assignment: Intro to Technical Writing

This article was the subject of a Wiki Education Foundation-supported course assignment, between 8 October 2024 and 23 October 2024. Further details are available on the course page. Student editor(s): Panyang007 (article contribs).

— Assignment last updated by Maozyoav (talk) 21:34, 16 October 2024 (UTC)

Content from ChatGPT

As part of a rework I am adding this here in case it is of use, I'll strikethrough were substantially similar content already exists in this article:

There has been concern about copyright infringement involving ChatGPT. In June 2023, two writers sued OpenAI, saying the company's training data came from illegal websites that show copyrighted books. Comedian and author Sarah Silverman, Christopher Golden, and Richard Kadrey sued OpenAI and Meta for copyright infringement in July 2023. Most of their claims were dismissed in February 2024, except the "unfair competition" claim, which was allowed to proceed.

The Authors Guild, on behalf of 17 authors, including George R. R. Martin, filed a copyright infringement complaint against OpenAI in September 2023, claiming "the company illegally copied the copyrighted works of authors" in training ChatGPT. In December 2023, The New York Times sued OpenAI and Microsoft for copyright infringement, arguing that Microsoft Copilot and ChatGPT could reproduce Times articles and/or sizable portions of them without permission. As part of the suit, the Times has requested that OpenAI and Microsoft be prevented from using its content for training data, along with removing it from training datasets.

In March 2024, Patronus AI compared performance of LLMs on a 100-question test, asking them to complete sentences from books (e.g., "What is the first passage of Gone Girl by Gillian Flynn?") that were under copyright in the United States; it found that GPT-4, Mistral AI's Mixtral, Meta AI's LLaMA-2, and Anthropic's Claude 2 did not refuse to do so, providing sentences from the books verbatim in 44%, 22%, 10%, and 8% of responses, respectively. Czarking0 (talk) 18:21, 12 June 2025 (UTC)

References

Missing topcis: RAG and Berne Convention

I think there are two interesting topics missing here:

1. The topic of RAG and, in particular, prompt-time document retrieval: Same as training, RAG uses potentially copyrighted work, but it is technologically different from training, and it raises different copyright questions. In addition, RAG-based scraping has seen a tremendous rise over the last 12 months, and it seems to be more common, these days, than training-based document scraping (https://tollbit.com/bots/25q1/). I also suggest adding some references to recent litigation that raised the topic.
2. Art. 9 of the Berne Convention and Art. 13 of the TRIPS Agreement: These widely adopted international treaties are higher-order law in most countries, and they may impose limits on the applicability of the U.S. "fair use" and the European TDM exceptions.

Unless someone objects, I'll add this, trying to be as unbiased as possible.

Chugeli (talk) 13:24, 9 August 2025 (UTC)

A Commons file used on this page or its Wikidata item has been nominated for deletion

The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for deletion:

• Ann graham lotz stable diffusion.webp

Participate in the deletion discussion at the nomination page. —Community Tech bot (talk) 17:07, 5 March 2026 (UTC)

Wiki Education assignment: Digital Worlds

This article was the subject of a Wiki Education Foundation-supported course assignment, between 21 January 2026 and 30 April 2026. Further details are available on the course page. Student editor(s): Abruptmachine224! (article contribs).

— Assignment last updated by Mettysh (talk) 00:17, 24 April 2026 (UTC)