Bengaluru, India – In early March, a realistic, AI-generated image of Indian Prime Minister Narendra Modi, styled as Bhishma Pitamah from the ancient Hindu epic Mahabharata, was promoted as a political advertisement on Instagram.
With long, wavy gray hair, a sun-shaped mark on his forehead and wearing body armor, the image depicted what Modi fans see as his role in today’s world: a reincarnation of the commander-in-chief Bhishma, who fought against foreign threats.
This Instagram image, created by right-wing Hokage Modi Sama and first posted in 2023, was promoted as a political ad for two days in March and received more than 35,000 impressions.
A review of Meta Ad Library data of political advertising in India over the past three months by Al Jazeera found that Hokage Modi Sama promoted nearly 50 AI-generated images of Modi between February 27 and March 21, making it the leading What has become advertisers is AI-generated Modi images on Instagram.
The Meta Ad Library is a public archive that hosts a collection of political advertising across its platforms, including Instagram and Facebook.
The common theme of all the images shared by the handle was the elevation of Modi as a Hindu leader. Popular AI images, amplified by sponsored posts on Hokage Modi Sama, depict Modi as the reincarnation of Bhishma, a son of God in a suit who embraces the Hindu heritage, and the king of Hindu Rashtra, ensconced on his throne, which Earned millions of likes and views. (Hindu Rashtra is the controversial ideology of Hindu majority rule in India, a departure from its secular founding principle.)
“Through these images, there appears to be an attempt to imbue Modi with a simultaneously wise and warrior quality, both of which create the aura of a political leader who is tireless, invincible, above reproach and therefore worthy of our unwavering loyalty,” Amogh Dar said Sharma, a lecturer at Oxford University who studies political communication, told Al Jazeera.
But the posts from the most prolific online advertiser of Modi’s AI images also highlight the challenges of enforcing AI-related rules on social media, amid fears that manipulated images could be used for propaganda among voters who are undermining the scale of the change that is coming Bring photos and photos, may not fully understand memes may have come through.
Aware of such use, Meta announced ahead of the crucial 2024 election year that, starting in January, political ads on Instagram and Facebook created with artificial intelligence (AI) will have to disclose the use or risk being banned.
According to data from Meta Ad Library, in the 30 days ended March 29, Hokage Modi Sama spent 537,799 Indian rupees ($6,500) promoting 363 pieces of political content, including images and videos, on his Instagram page. Our analysis shows that almost 14 percent of all sponsored ads, amounting to 50 images, were AI-generated.
All of Hokage Modi Sama’s AI ads were sponsored posts on Instagram, and disclosure of AI use was done through hashtags such as #aiartwork, #midjourneyart, #midjourneyai, and others. Midjourney, which the hashtags seem to refer to, is a popular generative artificial intelligence program.
But Meta told Al Jazeera that hashtags were not an acceptable disclosure for an advertisement that was created or modified digitally. In cases where advertisers are required to disclose that their content is digitally created or modified, Meta adds a “Digitally Created” label alongside the “Paid for by” disclaimer, including in the advertising library. Such labels are not currently present in Hokage Modi Sama AI advertisements.
Meta did not respond to Al Jazeera’s specific question about whether Modi’s photorealistic image ads violate his policy of depicting a real person doing something he didn’t do or depicting a realistic-looking event that didn’t happen.
Instead, Meta directed Al Jazeera’s reporter to a March 2024 blog post about AI disclosure requirements for advertisers and how the company is preparing for India’s elections.
Prateek Waghre, executive director of the India-based Internet Freedom Foundation (IFF), told Al Jazeera that content creators must acknowledge the attempt to promote AI images in any form, as these particular ads are tagged with #aiart be the disclosure to comply with Meta’s recently updated AI disclosure guidelines.
AI is used to drive “strategic narratives.”
Indian election campaigns have been a vehicle for a slow proliferation of AI-generated images and videos, including through AI resurrections of dead political leaders and their use as campaign art in official party reports, Al Jazeera previously reported. Despite the threat of AI-generated election misinformation, deepfakes are not used exclusively to deceive voters. Instead, generative AI is used to build a narrative.
“While disinformation is definitely a serious problem that requires our attention, it also diverts attention from the more subliminal ways in which AI-generated content can help advance strategic narratives of political parties,” Oxford University’s Sharma told Al Jazeera after checking the AI image advertising.
“It’s not just that AI ‘tricks’ voters into believing something that is obviously false; Rather, AI enables the production of content that is more creative and can draw on more innovative cultural references – resulting in political propaganda that is more entertaining and therefore more shareable, allowing for widespread distribution,” he added.
For example, the most popular AI image posted on the Hokage Modi Sama page is “Our Saffron Superhero.” It shows Modi wearing a saffron-colored kurta-pyjama and a flowing cape, walking beneath the “Om” flag, a Hindu spiritual symbol. The AI image has almost two million likes with comments like “Who wants India to be a Hindu nation?” in the comments section on Instagram. In most images, the Indian Prime Minister is portrayed as a “savior,” “saffron-clad guardian of India’s future,” “symbol of the Hindu renaissance,” and more.
Sharma pointed out that while mythologizing political leaders or depicting them in Hindu epics is nothing new, AI allows for a more sophisticated implementation.
“These AI-generated images of Narendra Modi appear to be the latest version of a long-standing strategy by the BJP to create a cult around Narendra Modi [Modi] as a wise and clever leader,” said Sharma.
While there are no explicit indications of direct affiliation with Modi’s Bharatiya Janata Party (BJP) on Hokage Modi Sama’s Instagram and Facebook pages, the page has published pro-Hindutva and pro-BJP content in the past. Previously, the page had gained popularity after sharing a Picture Elon Musk, the owner of the social media platform
Multiple attempts to contact Hokage Modi Sama’s administrator were unsuccessful. In the anime Naruto, “Hokage” is a prestigious title given to a village leader, while sama is a respectful way of saying “sir” in Japanese, so the page roughly translates to “Leader Modi Sir.” The Instagram account, which has 130,000 followers, now sells merchandise such as T-shirts, coffee mugs and diaries based on the Safran superhero AI image.
AI images have been used in political campaigns around the world. In March, supporters of US presidential candidate Donald Trump’s campaign used fake AI images of him posing with black voters to make “strategic outreach” to the black community. In Argentina, two presidential candidates used AI-generated images to boost their popularity and attack opponents. And in Indonesia, elected President Prabowo Subianto, the once feared military dictator, used AI images to portray himself as a “cuddly grandpa” during the election campaign.
In India, ahead of the Telangana state elections in December 2023, BOOM Live reported that AI images misrepresented regional political leader K Chandrashekar Rao at the launch of a free meal program for students, which he never attended.
A December 2023 review of Meta Ad Library by Al Jazeera found that the Mana Telangana page on Instagram and Facebook shared several fake AI images of Indian opposition leader Rahul Gandhi of the Congress Party posing with children and farmers. Farmers are an influential constituency in India and have frequently protested against Modi’s market-friendly farm laws.
Gandhi’s fake picture with the farmers was accompanied by the caption: “Rahul Gandhi empathizes with farmers and promises unwavering support and commitment to address their concerns individually.” The sponsored post included the hashtags #aigenerated and #aiimages, which suggesting they were AI-generated, and had an estimated 7,000 impressions.
West vs. Global South
There are also differences in the way platforms in the West and the Global South deal with the trend of AI imagery ahead of elections.
For example, to curb the potential misuse of AI images ahead of the US presidential election in November, popular AI image generator Midjourney banned the creation of fake images of President Joe Biden and Donald Trump in March.
However, there have been no such moves in India, raising “concerns about the fair enforcement of policies by tech platforms in different regions,” Baybars Orsek, the managing director of Logically Facts, told Al Jazeera. Logically Facts is part of Meta’s external fact-checking program and part of the Alliance to Combat Misinformation in India.
Al Jazeera’s attempts to create AI images of Modi and Gandhi shaking hands on Midjourney were successful, while a request to create an image of Modi and Trump shaking hands resulted in a “Forbidden prompt detected” notification .
“This decision brings to light the disparities in the application of such moderation measures around the world, particularly in the context of the Global South, including countries like India, which is facing the most significant election of the year,” Orsek said.
Al Jazeera emailed Midjourney about the disparity in moderation and has yet to receive a response.
“It is similar to the way social media companies like Facebook or Twitter bias their platform security policies toward the West, often neglecting perspectives from the global south,” IFF’s Waghre added.