Fake network using India's Covid-19 situation for pro-China propaganda: Report

The study suggests that the networks seek to influence several issues including US gun laws, human rights abuses in Xinjiang, and target the US for its response to India’s devastating second Covid-19 wave.

A network of more than 350 fake and repurposed social media accounts is pushing pro-China narratives and discrediting those that are critical of the Chinese government, according to a new study. Researchers at the Centre for Information Resilience (CIR) found that the narratives pushed by these fake or repurposed profiles on Twitter, Facebook and YouTube are similar to the content disseminated by Chinese government representatives and state-run media.

The network is using machine learning framework to create realistic-looking images of people who don’t exist. (CIR)

The study suggests that the network seeks to influence several issues including US gun laws, human rights abuses in Xinjiang, global conflict, racial discrimination and US response to India’s devastating second Covid-19 wave. Citing the grim situation in India during the second Covid wave, the network targets the response of the US.

Some of these social media accounts alleged that the US was allowing vaccines to expire when there was a dire need for Covid-19 shots in India. In a YouTube video titled ‘India's out-of-control epidemic highlights the hypocrisy of American human rights’, the social media account claimed that large stockpiles of unused vaccines were kept in factories in the United States when India needed those vaccines.

The video also features a tweet from Serum Institute of India (SII) CEO Adar Poonawalla in which he urged US President Joe Biden to lift the embargo on exports of raw material needed for vaccine manufacturing. “Some of the network has promoted long-text memes on Twitter about India and coronavirus with negative sentiment towards the US,” the report said.

Some of the Twitter handles involved in pro-China propaganda were found to be using a mix of machine-learning generated images as profile pictures and relying on a much larger network that was using more authentic appearing images, anime images and repurposed accounts. The AI-generated images are created by a type of machine learning framework called StyleGAN to create realistic-looking images of people who don’t exist.

The report also suggests that there was a change in ownership of authentic-appearing accounts at some point, either through a “takeover from password dumps or a purchase from a seller of stolen accounts.” The researchers conducted a manual review of hundreds of accounts across different social media platforms and concluded that there are between 300-500 such accounts on Twitter, 40-55 accounts and pages on Facebook and 12 accounts on YouTube.

Benjamin Strick, the author of the report and CIR’s director of investigations, said in a statement that the aim of the network appears to delegitimize the West by amplifying pro-Chinese narratives. He said that the content was posted in English and Chinese but there’s no concrete evidence that these accounts are linked to the Chinese government.

“The narratives CIR uncovered in this influence operation have strong similarities to content seen on the accounts of Chinese Government representatives and China state-linked media. However, this is not an indication of attribution,” he added.