The California legislature has passed a controversial bill-the AI Training Data Transparency Act-which will now be sent to Governor Gavin Newsom for approval. However, Governor Newsom has not yet expressed his stance on the bill.
The bill aims to mandate that artificial intelligence companies provide greater transparency regarding the datasets used in the development and training of their models. Initially introduced by Assemblymember Jacqui Irwin, the bill (AB 2013) originally had a broader scope, but on August 20, it was narrowed to apply specifically to generative AI-AI that can create text, images, and similar content, such as OpenAI's ChatGPT.
The legislation requires AI companies to disclose key information about their training models, including the sources of data, the types of data used, and how the data is processed. This is significant for protecting data privacy and preventing data misuse, as well as helping the public better understand AI technologies. However, industry stakeholders are concerned that overly stringent transparency requirements could negatively impact companies' trade secrets and stifle technological innovation.