SAN DIEGO—With concerns running high about deepfakes misleading voters during the upcoming election season, Verance has unveiled a watermarking system for digital authentication and tracking of trusted news sources of video.
The company’s watermarking system comes as concern grows among the public about the impact of deepfakes on the election. An Axios-Morning Consult poll conducted in August found more than half of Americans expect misinformation spread by AI to impact who wins.
Major news organizations are taking steps to bolster their vigilance against deepfakes. For instance, CBS News has recently launched a new fact-checking unit for AI to combat misinformation from deepfakes.
Verance currently is conducting advanced talks with TV and digital news operations and consumer electronics giants about creating partnerships for the trusted news source watermarking system. The watermarks embedded in video content would be detected by receivers in the home, enabling viewers to decide about the authenticity and accuracy of information, the company said.
The system will give viewers the ability to determine if the media they are watching has been tampered with or modified during distribution. It also equips them with access to fact-checking sources, the company said.
The watermarking system will be available to political campaigns, government and corporations as well. Working with manufacturers of TVs and other consumer devices, these entities can protect viewers from deepfakes and other misinformation, it said.
“It is important to enlarge the body of authenticated news content to reduce the influence of content that could be fake, especially in a hotly contested election year,” said Verance CEO Nil Shah.
“Without universally trusted authenticity tools in this new era, the internet’s societal benefit as a means of accessing knowledge will precipitously decline. This presents a serious problem not only for large technology platforms such as Facebook and YouTube but also traditional news media that increasingly rely on the internet for sourcing and verification.”
The system is in line with an Executive Order President Joseph Biden released Oct. 30 that builds on voluntary commitments already made by technology companies, the company said.
The order requires industry to develop safety and security standards before AI products can be introduced to the public and gives federal agencies a to-do list to oversee AI deployment. It directs the Department of Commerce to issue guidance to label and watermark AI-generated content to help differentiate between authentic interaction and those generated by software.
Deadlines set out in the order range from 90 to 365 days for implementation. First up are safety and security items.
More information is available on the company’s website.
