How does OpenAI claim its AI training data is lawful?

Question

Accepted Answer

OpenAI Asserts Lawfulness of AI Training Data Based on Public Availability and Fair Use Principles

OpenAI asserts that its artificial intelligence models are trained on data that is publicly available. The company states that this training is underpinned by principles of fair use and relevant international copyright laws, which it contends are equitable for content creators. Furthermore, OpenAI highlights its collaborative efforts with news publishers, which include displaying their content, providing attribution, and linking to it within ChatGPT's search functions. For publishers who wish to prevent their data from being used, OpenAI offers straightforward opt-out mechanisms.
Training Data Sourced from Publicly Available Information

OpenAI's defence against claims of copyright infringement centres on the assertion that the data used to train its AI models was sourced from publicly accessible platforms. This approach, the company argues, aligns with established legal frameworks for data utilisation in technological development.
Reliance on Fair Use and International Copyright Principles

The company's legal stance is grounded in the concept of fair use, a doctrine that permits the limited use of copyrighted material without requiring permission from the rights holders for purposes such as criticism, comment, news reporting, teaching, scholarship, or research. OpenAI extends this argument to include related international copyright principles, positing that these frameworks are designed to be fair to creators while enabling technological advancement.
Collaboration and Attribution with News Publishers

OpenAI maintains that it actively collaborates with news organisations. This collaboration reportedly involves integrating publishers' content into its services, such as displaying articles, providing proper attribution to the source, and including direct links to the original content within its ChatGPT search feature. Such practices, OpenAI suggests, demonstrate a commitment to respecting publishers' rights.
Publisher Opt-Out Mechanisms Provided

Recognising the concerns of content owners, OpenAI has stated that it provides publishers with accessible methods to opt out of having their data used for AI model training. This mechanism is presented as a straightforward way for news media companies and other creators to control the use of their intellectual property by the company.

Reading history

Hide →

This history is cleared when the page is closed.

How does OpenAI claim its AI training data is lawful?

Loading

Was this article helpful?

Helpful

Not helpful

Report an issue

1

0

0

Share with friends