Yahoo News has signed up to use San Jose-based cybersecurity company McAfee’s deepfake image detection technology. The scalable McAfee system can “quickly identify images that may have been produced or modified using AI, including deepfake images,” flagging them for the Yahoo News editorial standards team for human review. The standards team then “determines whether the flagged images meet the platform’s editorial guidelines.” The partnership provides news aggregator Yahoo with an extra layer of protection as it deals with a large network of global publishers in addition to policing its original content.

Comscore consistently ranks Yahoo — owned since 2021 by investment firm Apollo Global Management — among the top 5 sites with the most unique U.S. visitors.

In a news release McAfee writes the deal “integrates McAfee’s deepfake image detection technology into the robust content quality system on Yahoo News,” providing the editors with “deeper insights into whether AI has been used to create and modify images” while allowing both to help maintain “high-quality, credible journalism.”

McAfee plans to continually update its system and the models that power it as it monitors changes in the techniques used to alter images with AI, according to an interview with CTO Steve Grobman in Variety. The system McAfee is putting in place will “provide updates automatically to the process that Yahoo News is using,” Grobman said.

Forbes explains the system, “powered by McAfee Smart AI, uses advanced machine learning algorithms to identify inconsistencies typical of AI-generated content.” On a granular level, it works by “analyzing the unique patterns left behind when AI generates or alters an image.”

“These patterns, while often undetectable to the naked eye, can be identified by AI models trained to spot them,” explains Forbes. “The tool then flags the image for review, where it can be cross-referenced against other known sources or scrutinized for further signs of manipulation.”

The Los Angeles Times provides examples of audio deepfakes circulating on X. Though there’s no word as to McAfee developing a technology to detect audio and video deepfakes, it seems likely development is underway, as Grobman told LA Times that “one of the really neat things about AI is, you don’t need to tell the model what to look for. The model figures out what to look for.”

Others are also working on guardrails. Variety points out that the Publicis Groupe’s agency, Team One, has introduced free app called Faikcheck “that can help marketers scan for deepfake spoofs and images of their brands and ad messages.”

Drexel University and Intel, with FakeCatcher, have also touted their development of accurate detection tech.