China has started using AI in influence operations, Microsoft says

Suspected Chinese disinformation operatives are already starting to leverage generative artificial intelligence in their influence campaigns, Microsoft warned Thursday.

Driving the news: Microsoft said in a new report that it has uncovered evidence that suspected China-linked actors are using AI-generated images in their social media campaigns to make them more believable.

• China-linked influence operations have been using AI-generated images since at least March, Microsoft said, and the images are already drawing "higher levels of engagement from authentic social media users."

Why it matters: This is one of the first known examples of a nation-state abusing generative AI in its cyber operations.

Details: Microsoft noticed suspected Chinese operatives using AI-generated images in a monthslong campaign that focused on both divisive U.S. political topics, like gun violence, and "denigrating U.S. political figures and symbols."

• One example of an AI-generated image involves the Statue of Liberty holding a rifle as well as the customary torch, with the text "The Goddess of Violence" overlaying it.
• However, the images still have some flaws: In the Statue of Liberty example, the image gave the statue more than five fingers on one of its hands.

The big picture: Cybersecurity and national security officials have long speculated that adversaries would tap generative AI to make phishing and disinformation campaigns more believable.

What they're saying: "We can expect China to continue to hone this technology over time, though it remains to be seen how and when it will deploy it at scale," Clint Watts, general manager at Microsoft Threat Analysis Center, wrote in a blog post.

The intrigue: Typically, China-linked influence operations have struggled to garner authentic engagement across social media.

• But Microsoft has found that, broadly speaking, China-affiliated influence operations are achieving more on social media as they lean on generative AI, as well as real people, to create tailored, authentic content.

Sign up for Axios' cybersecurity newsletter Codebook here