Litigation against humanity over alleged fraudulent data scraping

Reddit has launched a lawsuit against humanity for artificial intelligence companies, accusing the company of using Reddit content without permission to slash the platform and train Claude AI models.

Complaint, Submit On Wednesday, a US federal court alleges that humanity continued to access Reddit servers, including more than 100,000 times, after publicly alleging that it breached Reddit user agreement and ceased such activities in July 2024.

Reddit seeks damages, compensation, and court orders that prohibit humanity from using Reddit-derived data in its products, including preventing the license or benefits of AI programs trained with the content of Reddit content.

Decryption We contacted humanity about our response to Reddit’s claims.

The social media giant claimed that AI companies have “two faces.” This sought to establish himself as a responsible player in the AI ​​industry.

“There are public faces that try to infiltrate consumer consciousness with the claims of ((borderlines and law justice and respect), and private faces that ignore rules that prevent attempts to line up more pockets,” the lawsuit reads.

At the heart of the conflict is the widespread controversy over the massive training of language models. Since Openai’s ChatGPT debut, concerns have been growing over the use of both copyrighted and user-generated materials in AI development.

Ongoing Issues

Several lawsuits have already been filed by the organization. New York Times To Openai and Microsoft in 2023. Other plaintiffs include visual artists, authors and record labels that claim that their work was exploited without permission.

So does humanity surface Another lawsuit relating to alleged use of lyrics to copyrighted songs and another lawsuit from a group of authors who said the company was using pirated versions of their books as training material.

The tension ripples over the cultural setting, with artists expressing their anger at the Ai-generated imitation of their style.

Earlier this year, Replica The art style of popular Japanese animation company Studio Ghibli has sparked copyright violations and concerns that artists have lost to AI programs trained with their works.

Last year, it was submitted to the UK Parliament, and Openai Recognised Using copyrighted content in training and claiming that without it it is “impossible” to develop major AI systems. The company argues that such practices are legal.

A proposal in the UK last month Easy Copyright law and the use of copyrighted material to train LLMS has been fired from prominent artists, including Elton John.

However, despite the protests over the protection of users, Reddit itself doesn’t think it’s very wrong to use user content for LLM training, as long as Reddit is compensated.

We have signed our own licensing agreements with companies such as Openai, Google, Sprinklr, and Cision to allow access to content for training purposes.

edit Sebastian Sinclair