Russia is using generative artificial intelligence (AI) in online deception campaigns, but the efforts have so far been unsuccessful, according to a security report published Thursday by Meta.

The parent company of Facebook and Instagram found that AI-powered tactics have so far delivered “only incremental productivity and content generation gains” for bad actors and that Meta has been able to disrupt deceptive influence operations.

Meta’s efforts to combat “coordinated inauthentic behavior” on its platforms come amid growing fears that generative AI will be used to mislead or confuse people in elections in the United States and other countries.

Russia remains the largest source of “coordinated, inauthentic behavior” through fake Facebook and Instagram accounts, Meta’s director of security policy David Agranovich told reporters.

Since Russia’s invasion of Ukraine in 2022, those efforts have been primarily aimed at undermining Ukraine and its allies, the report said.

As the US elections approach, Meta expects to see Russian-backed online deceptive campaigns targeting political candidates who support Ukraine.

Facebook has been accused for years of being a powerful platform for election disinformation. Russian agents used Facebook and other social media platforms in the U.S. to stoke political tensions in multiple U.S. elections, including the 2016 election won by Donald Trump.

Experts fear an unprecedented flood of disinformation from malicious actors on social networks due to the ease with which generative AI tools such as ChatGPT or the Dall-E image generator can create on-demand content in seconds.

According to the report, AI is being used to create images and videos, and to translate or generate text, and to create fake news stories or summaries.

When Meta looks for cheating, it looks at the behavior of accounts rather than the content they post.

Influencer campaigns often span a range of online platforms, and Meta has noted that posts on X, formerly Twitter, were used to make fabricated content appear more credible. Meta is sharing its findings with X and other internet companies, saying a coordinated defense is needed to counter misinformation.

“As for Twitter (X), they’re still in transition,” Agranovich said when asked if Meta thinks X will act on deception tips. “A lot of the people we’ve dealt with in the past have moved on.”

X has undermined trust and safety teams and scaled back content moderation efforts once used to contain misinformation. Researchers are now calling it a breeding ground for disinformation.