Microsoft sees slowdown in Russia as China struggles to influence US vote

(Bloomberg) — Russia is working to undermine U.S. support for Ukraine ahead of the 2024 elections, while Chinese influence campaigns aim to divide Americans and undermine their faith in democracy, according to new research from Microsoft Corp.

Bloomberg’s Most Read

But both countries’ persistent efforts to shape American politics through propaganda and cyberattacks were less pronounced during this year’s election cycle than in 2016 and 2020, the company said in a report published Wednesday. It also found that widespread fears that voters would be influenced by deepfake videos made with artificial intelligence have not yet come to pass, although simpler efforts have been more persuasive.

Representatives from the Chinese and Russian embassies in Washington did not immediately respond to requests for comment.

Microsoft researchers found several “Russian-affiliated” efforts to promote disinformation to harm Ukrainian President Volodymyr Zelenskiy or cause Americans to question his government’s support for Ukraine as it continues to combat Russia’s invasion in 2022. These Campaigns have been conducted in English and Spanish and often take the form of alleged whistleblowers or amateur journalists publishing misinformation that is then covered by apparently unaffiliated websites, the report states.

Microsoft found that the most prolific promoters of these narratives “are supported by or affiliated with” the administration of Russian President Vladimir Putin and that the disinformation is assumed by Americans who are likely unaware of its source. The company also said it has seen an increase in recent months of a hacking group affiliated with the Russian military targeting Western policy think tanks.

China has made greater use of AI to create and enhance images, memes and videos aimed at stoking controversy in the US and deepening existing social divisions, Microsoft has found. For example, social media accounts linked to China have spread conspiratorial content, such as posts featuring AI-generated images of burning coastal roads that blame the US government for the summer 2023 wildfires on the Hawaiian island of Maui.

The researchers concluded that this type of relatively simple digital forgery attracted attention, although there was little sign of people being widely influenced by sophisticated AI-made deepfake videos.

The story continues

“Rarely has the deployment of AI-enabled generative content by nation-states achieved such reach on social media, and only in a few cases have we seen any genuine deception of the public with this content,” the report states.

It also concluded that faked audio has been more effective as propaganda, as it is easier to make and more difficult to debunk than video.

This type of deception caused alarm in January when a doctored audio message from US President Joe Biden circulated discouraging voters in New Hampshire from voting in the state’s Democratic primary. In this case, the audio was not attributed to a foreign government, but rather to an American political agent.

Bloomberg Businessweek Most Read

©2024 Bloomberg LP