Commentary

High Watermark: OpenAI And Microsoft Support Bill To Enforce AI Transparency

California is mulling yet another bill to regulate the tech/publishing business. Only this time some of the chief companies support the measure.  

OpenAI has signalled its backing of AB 3211, a bill that would require companies to use watermarks in metadata for AI-powered photos, audio and videos.

This bill, which is on its way to a vote, is designed to combat the kind of disinformation that may spread during this election year.   

"New technology and standards can help people understand the origin of content they find online, and avoid confusion between human-generated and photorealistic AI-generated content," wrote Jason Kwon, chief stratrgy officer at OpenAI, according to Reuters.  

Microsoft and Adobe also support AB 3211, TechCrunch reports.

As written, the bill would authorize the state Department of Technology to prescribe penalties of up to $100,000 for each intentional or negligent violation.  

advertisement

advertisement

This sounds worthy, although the premise seems to be that all AI-generated visual content is suspect. This may be a piece of over-regulation that publishers will find burdensome. However, many others are already using watermarks as a sort of standard best practice.

There may come a day when genAI material is more widely accepted. What do we do then? 

The bill’s co-sponsor is Assemblyperson Buffy Wicks (D), who spearheaded the California Journalism Preservation Act. That piece of legislation was rendered moot when the state and Google reached an agreement that will result in $250 million in funding for newsrooms.  

Some journalists object that the bill is a giveaway for Google. But Wicks has defended it, saying, in effect, that something is better than nothing. 

Instead of simply trashing the bill, they should write a bill that reflects the agreement with Google. Wicks should have her name on such an accomplishment.
Next story loading loading..