AI-generated images and videos are a significant threat to proper and accurate information. Fake imagery causes confusion, panic, hate, and bullying, and it’s now easier to create than ever. But there’s a step forward to resolving, or at least minimizing, the issue. Seven major tech companies, including OpenAI, Microsoft, and Google, have promised to create watermarks for AI-generated content. This way, it could become safer to share AI content without misleading people about its authenticity.
[Related reading: Bye-bye fake news: this free tool spots AI-generated images in a second]
Other than OpenAI, Microsoft, and Google, the companies joining the effort are Meta, Amazon, Anthropic, and Inflection. They have promised to label the AI-generated content, but for now, it’s not exactly clear how they will do this watermarking will. However, we can assume that it will be integrated with the content. This way, people can trace where it came from and what AI tools were used to make it. If you’ve played with Adobe Firefly, you’ve noticed that its images already have a watermark like that.
The White House believes that “the watermark will enable creativity with AI to flourish but reduces the dangers of fraud and deception.” It would be a win-win situation, and tech companies have responded to the concerns and suggestions.
OpenAI responded to the White House’s concerns, committing to manage various risks and concerns. They include safety, security, and trust, among other things. OpenAI writes that it will “develop and deploy mechanisms that enable users to understand if audio or visual content is AI-generated, including robust provenance, watermarking, or both, for AI-generated audio or visual content.”
Google has also addressed the potential issues with AI-generated content and the solutions it plans to incorporate. The company plans to include metadata and other techniques, along with watermarking, to encourage reliable information. Google added, however, that no single company can manage AI correctly alone.
“When it comes to content, we’re taking steps to promote trustworthy information. We’ll soon be integrating watermarking, metadata, and other innovative techniques into our latest generative models, and bringing an About this image tool to Google Search to give you context about where an image first appeared online. Addressing AI-generated content will require industry-wide solutions, and we look forward to working with others, including the Partnership for AI’s synthetic media working group.”
Microsoft has also committed to developing a “safe, secure, and trustworthy AI.” They don’t explicitly mention watermarking, though, neither in the blog post nor in the more detailed AI managing plan. the company greeted the government’s efforts in balancing the potential and risks of AI and encouraged the tech industry to make AI safer and more beneficial.
Safe use of AI should be a joint effort
In addition to watermarks, tech companies have also voluntarily agreed to test AI systems before their release, invest more in cybersecurity, and share information to decrease AI risks. Microsoft has also promised to support the creation of a national registry to keep track of high-risk AI systems.
But as Google said, “None of us can get AI right on our own.” And it’s not just major tech companies, but I’d also extend this to individuals who use and consume AI-generated content. I believe that we can all have a role in this new and fast-changing industry and the way we use it. As producers, we need to be clear about whether or not content is AI-generated and if it contains accurate information. And as consumers, we need to develop critical thinking and reasonable doubt instead of believing everything we see – now more than ever. We can also check whether images are AI or not, which I think is awesome. But having tech companies and major news outlets be transparent about it is certainly a must. As individual creators and consumers, we can’t manage AI on our own, either.