Use Of Artificial Intelligence In Online Influence Campaign Increases, Including In Indonesia

JAKARTA - The cybersecurity firm Mandiant owned by Google in the United States said on Thursday 17 August that it had seen increased use of artificial intelligence (AI) to conduct online manipulative information campaigns in recent years, although the use of this technology in other digital intrusions was still limited.

Researchers at the Virginia-based company have found "many examples" since 2019 where content generated by AI, such as a fabricated profile picture, has been used in a politically motivated online influence campaign.

This includes campaigns from groups in line with governments of Russia, China, Iran, Ethiopia, Indonesia, Cuba, Argentina, Mexico, Ecuador and El Salvador, the report said.

This has happened amid recent booms in generative AI models like ChatGPT, which make it easier to make fake videos, images, text, and computer codes convincing. Security officials have warned about such models used by cybercriminals.

Generative AI will allow groups with limited resources to generate higher-quality content on a larger scale, researcher Mandiant said.

"For example, a information campaign related to a Chinese country called Dragonbridge, has developed "exponentials" through 30 social platforms and 10 different languages since it first started targeting pro-democracy demonstrators in Hong Kong in 2019," said Sandra Joyce, vice president of Mandiant Intelligence.

However, the impact of these campaigns is limited. "From the point of view of effectiveness, there aren't many wins there," Joycem was quoted as saying by Reuters. "They've really haven't changed the direction of the threat landscape so far."

China has denied US accusations of involvement in such an influence campaign before.

Mandiant, which helps public and private organizations respond to digital abuses, said they had not seen AI play a key role in threats from Russia, Iran, China, or North Korea.

"So far, we haven't seen a single response incident in which AI plays a role," Joyce said. "They haven't really been used in the type of practical use that is superior to what can be done with normal tools we've seen."

But he added: "We can be very confident that this will be a bigger problem over time."