AI-Powered Social Media Accounts Spread Indian Army Propaganda

A large network of fake social media accounts promoting Indian government and army propaganda has been uncovered after operating for nearly three years. The network, which primarily targeted Indian audiences, was spreading content favourable to Prime Minister Narendra Modi’s administration while casting negative portrayals of neighbouring countries like China, the Maldives, Bangladesh, and Pakistan.

Researchers from NewsGuard, a group specializing in identifying misinformation, revealed that this vast operation involved at least 500 Facebook accounts and 904 X (formerly Twitter) accounts. These accounts actively posted, reposted, and commented on pro-government content, often creating the illusion of widespread support for India’s military and political leadership. The network’s primary goal seemed to be to bolster public favour for Modi’s administration and Indian military efforts while discrediting other countries, particularly Pakistan and China.

AI-Powered Social Media Accounts Spread Indian Army Propaganda

According to Dimitris Dimitriadis, the Director of Research and Development at NewsGuard, the discovery of this network was unexpected, especially given how long it had remained undetected. “We regularly track inauthentic networks, but they typically get taken down within weeks,” said Dimitriadis. This campaign, however, had managed to evade detection since September 2021, making it one of the longest-running influence operations NewsGuard has come across.

Tactics and Content

Despite managing to stay hidden for so long, the operation was far from sophisticated. The accounts used fake names and profile pictures and often posted content that was bland and repetitive. Rather than spreading outright disinformation, the network primarily promoted pro-government propaganda by reposting favourable news stories from pro-government outlets like the Hindustan Times and ANI News.

For instance, in July, 20 fake X accounts tied to the network commented on a post from ANI News about General Upendra Dwivedi becoming the new Chief of Army Staff. The fake accounts added nearly identical comments such as, “The Indian Army — A symbol of national strength” and “General Dwivedi — A leader who values transparency and accountability.” The creators designed these cookie-cutter responses to project a unified show of support for the Indian military.

Some accounts, like the ironically named JK News Network, purported to offer 24/7 news updates. Instead, these accounts predominantly posted pro-military content and flattering photos of military personnel. Often, the text accompanying these posts appeared AI-generated, filled with awkward phrasing, sloppy English, and incomplete sentences, indicating that the operation lacked human oversight.

Repetition and AI-Generated Content

The network’s use of repetition was another clear indicator of inauthentic behaviour. Many of the same accounts posted identical content multiple times a day. For example, 429 other fake accounts reposted the message verbatim when JK News Network posted about Pakistan’s treatment of religious minorities in Balochistan.

Despite the campaign’s wide reach, it appeared to have little impact. Dimitriadis explained that such influence operations often struggle to gain traction because they fail to engage with real public conversations. He described the campaign as “awkward” and “sloppy,” which likely contributed to its lack of engagement.

Lack of Detection

One reason the campaign went undetected for so long might be the platforms themselves. According to Abu Qureshi, a threat intelligence expert at BforeAI, social media algorithms tend to display content from just a few accounts to gauge user engagement. This makes it easier for fake accounts to avoid detection, especially when they target like-minded individuals who are unlikely to report or flag suspicious content.

The discovery of this network serves as a reminder of the ongoing threat of online influence operations and the need for vigilance when consuming social media content.

See Also: IHC Takes Action Over Social Media Campaign Targeting Judges