Reddit Takes Legal Action Against Anthropic for Unauthorized Use of Data

Reddit has initiated a lawsuit against the AI startup Anthropic, claiming that it allegedly utilized data from the site to train its models without obtaining the necessary licensing agreement.

The company argues that using this information for commercial purposes violates the user agreement of the social platform and constitutes a legal infringement.

Sam Altman is one of Reddit’s major shareholders; he invested in the company back in 2014 and currently holds an 8.7% stake. He ranks as the third-largest shareholder, behind Advance and Tencent.

Reddit is the first among major tech firms to legally challenge AI developers regarding the training of neural networks. Similar legal actions have occurred in the past.

The New York Times has filed a lawsuit against OpenAI and Microsoft for tuning AI models based on its articles without compensation or permission. A similar lawsuit has been brought against Meta by authors Sarah Silverman, Christopher Golden, and Richard Kadri, alleging that the company used their books to train a language model, infringing on copyright.

In May 2024, eight daily newspapers owned by Alden Global Capital also sued OpenAI and Microsoft, accusing them of illegally using «millions of articles» to train AI products without permission.

“We will not tolerate profit-driven entities like Anthropic exploiting Reddit’s content for billion-dollar revenues while offering nothing in return to our platform’s users and showing a lack of respect for their privacy,” stated Reddit’s chief legal officer, Ben Lee.

Interestingly, Reddit has made agreements with other AI model providers, including OpenAI and Google, allowing them to train neural networks on its data. These deals are said to include provisions to safeguard user interests and privacy.

“We disagree with Reddit’s claims and will vigorously defend ourselves,” commented Anthropic representative Daniel Giggliere.

In April 2023, it was revealed that Reddit planned to start charging companies for access to its API for gathering data to train large language models.

Reddit is seeking damages from Anthropic for the alleged harm caused and reimbursement for the profits the startup gained through its use of the platform’s content. The company is also pursuing a court order to halt any further use of its information.

Previously, Anthropic launched a blog for its chatbot, Claude, featuring posts on technical subjects related to various applications of the digital assistant.