Rising AI Use in Online Influence Campaigns, Limited Impact: US Cybersecurity Firm

On August 17, the Google-owned U.S. cybersecurity firm Mandiant reported that it has observed a rising trend in the use of artificial intelligence (AI) for manipulative online information campaigns. Although the application of this technology in other digital intrusions has been somewhat restricted, researchers from the Virginia-based company identified multiple cases since 2019 where AI-generated content, including fake profile pictures, was employed in politically-driven online influence campaigns.

The report noted that these campaigns involved various groups associated with the governments of Russia, China, Iran, Ethiopia, Indonesia, Cuba, Argentina, Mexico, Ecuador, and El Salvador.

This development coincides with a recent surge in the popularity of generative AI models like ChatGPT, which facilitate the creation of persuasive fake content encompassing videos, images, text, and computer code. Security authorities have raised concerns about the potential exploitation of such models by cybercriminals.

The researchers at Mandiant pointed out that the adoption of generative AI could empower groups with constrained resources to create more sophisticated content for influence campaigns on a larger scale.

For instance, a pro-China information campaign known as Dragonbridge witnessed significant growth, spreading "exponentially" across 30 social platforms and spanning 10 different languages. This expansion occurred after the campaign's initial focus on targeting pro-democracy activists in Hong Kong back in 2019. Sandra Joyce, Vice President at Mandiant Intelligence, provided this insight.

However, the overall impact of these campaigns was found to be limited. Sandra Joyce stated, "From an effectiveness standpoint, not a lot of wins there. They really haven't changed the course of the threat landscape just yet."

China has consistently refuted previous U.S. allegations of involvement in such influence campaigns.

Mandiant, a firm that assists both public and private organizations in responding to digital breaches, noted that AI has not yet played a significant role in threats originating from Russia, Iran, China, or North Korea. The researchers anticipate that the use of AI for digital intrusions will remain relatively low in the near future.

Sandra Joyce emphasized, "So far, we haven't encountered any incident response where AI had a role. They haven't been practically employed to a degree that surpasses what conventional tools can achieve."

Nevertheless, she concluded, "We can confidently predict that this issue will escalate over time."