A 6578 — Establishes The Artificial Intelligence Training Data Transparency Act
A 6578 establishes the Artificial Intelligence Training Data Transparency Act, which aims to increase transparency around the data used to train AI systems. The bill requires AI developers to disclose the sources and characteristics of training data, particularly when such data may lead to biased or discriminatory outcomes. It addresses harms related to algorithmic discrimination and privacy risks associated with opaque AI systems. The bill was amended on third reading on February 24, 2026, and remains under review.
Linked Incidents
3Incidents this policy has been directly linked to
New York City business owners receive illegal advice from MyCity AI chatbot, leading to bot shutdown in early 2026
Middle-aged man wrongfully arrested after AI misidentification in criminal investigation
Seattle resident banned from Madison Square Garden venue after t-shirt design linked via facial recognition to 2021 incident he did not attend
Related Incidents
Same harm domain, actors and location may differ
Punjab Kings cricketer targeted by AI-generated deepfake video leading to reputational harm and public confusion
Adult film actors fann wong and christopher lee defrauded via ai-generated forged images
Veteran actress Yeom Hye Ran defrauded of likeness via unauthorized AI-generated video on YouTube
Renowned South Korean actress has AI deepfake portrait rights violated through unauthorized image generation technology
5 million Abceed users exposed to AI voice cloning fraud risk via misconfigured cloud storage
Related Legislation
Other policies covering the same harm domain