Major tech companies are voicing support for new legislation in California aimed at increasing transparency around artificial intelligence generated media. The bill, called AB 3211, proposes requirements for labeling synthetic content and attaching provenance data.
If passed, the proposed law would mandate the inclusion of metadata on images, videos and audio files created using AI. This provenance data must specify whether the content is computer generated or human made, along with details like the time of creation and responsible developer. The goal is to reduce confusion and avoid manipulated AI content being presented as genuine.
Supporters of the bill argue it can help the public identify synthetic media and promote trust online. Companies backing AB 3211 include Microsoft, Adobe and ChatGPT creator OpenAI. As leaders in the AI industry, their endorsement provides credibility that the regulations aim to address valid concerns, not stifle innovation.
The bill also wants to establish tools for reviewing embedded provenance data and requirements that such markings cannot be easily stripped from files. This addresses the challenge that metadata can currently be removed, making artificial content harder to detect.
While initially opposed by some industry groups, amendments to lessen penalties for inadvertent violations seem to have gained Microsoft and Adobe’s approval. If approved, the new standards could set an example for managing advanced generative models and the content they produce, at a pivotal time as such technologies continue advancing.
Backers hope transparent sourcing demands will help the public identify authentic information online, especially in the lead up to major elections. With further support and responsible oversight, the bill aims to strike a balance between openness and appropriate oversight of increasingly sophisticated AI.
