New York Senate Bill S06955

New York 2025-2026 Regular Session

• Introduced: 3/27/25
• Refer: 3/27/25
• Refer: 1/7/26

Caption

Establishes the artificial intelligence training data transparency act requiring developers of generative artificial intelligence models or services to post on the developer's website information regarding the data used by the developer to train the generative artificial intelligence model or service, including a high-level summary of the datasets used in the development of such system or service.

Impact

If enacted, the bill would significantly impact how AI developers operate within New York State, adjusting the legal landscape to require greater accountability over AI data sources and usage. This development aligns with broader trends towards regulatory frameworks for AI technologies, reflecting growing public concern over data privacy and ethical AI practices. The act would hold developers responsible for ensuring that data used in AI training does not infringe on privacy rights and that any associated risks are clearly communicated to users.

Summary

Bill S06955, also known as the Artificial Intelligence Training Data Transparency Act, aims to enhance transparency around the datasets used for training generative artificial intelligence models or services. This legislation mandates that developers must publicly disclose specific information regarding the data used in their AI models. The required disclosures include details on data sources, the number of data points, cleaning processes, and the presence of personal information within the datasets. These measures aim to inform users and stakeholders about the underpinnings of AI technology, thereby fostering greater accountability in the deployment of AI services.

Contention

There are notable points of contention surrounding Bill S06955. Supporters argue that the transparency requirements will promote ethical AI development and mitigate risks associated with data misuse. Critics, however, express concerns that such regulations might impose excessive burdens on smaller developers, potentially stifling innovation. Additionally, debates may arise about how the terms used in the bill—such as 'personal information'—are defined, which could influence the extent of compliance required by small entities versus larger corporations. As the bill moves through the legislative process, these discussions will be crucial in shaping its final form and implementation.