NEW DELHI – Bollywood stars rarely get entangled in politics, so videos showing two celebrities criticizing Indian Prime Minister Narendra Modi – and supporting his fundamental opposition, the Congress Party – were certain to go viral.
But the clips of A-list actors Aamir Khan and Ranveer Singh were fake, AI-generated videos that were one other example of the false or misleading claims circulating on the web with the aim of influencing India's elections. Both actors filed police reports, but such measures do little to stem the flow of such misinformation.
Recent claims circulating online in India include misrepresentations about voting, claim without evidence that the election was rigged, and incite violence against India's Muslims.
Researchers tracking misinformation and hate speech in India say tech firms' poor enforcement of their very own policies has created perfect conditions for harmful content that would distort public opinion, incite violence and leave thousands and thousands of voters wondering what to consider.
“A non-sophisticated user or normal user has no idea whether it is someone, an individual, sharing his thoughts on the other end or whether it is a bot?” Rekha Singh, a 49-year-old voter, told The Associated Press. Singh said she fears social media algorithms are distorting voters' view of reality. “So they are biased without even realizing it,” she said.
In a yr filled with major elections, the massive variety of elections in India stands out. The world's most populous country has dozens of languages, the biggest variety of WhatsApp users and the biggest variety of YouTube subscribers. Nearly a billion voters are eligible to solid their ballots within the election, which runs into June.
Tech firms including Google and Meta, the owner of Facebook, WhatsApp and Instagram, say they’re working to combat misleading or hateful content and help voters find reliable sources. But researchers who’ve long tracked disinformation in India say their guarantees ring hole after years of failed enforcement and “coordinated” approaches that fail to have in mind India's linguistic, religious, geographical and cultural diversity.
Given India's size and importance to social media firms, one might expect a greater focus, say disinformation researchers who give attention to India.
“This is how the platforms make money. They benefit and the whole country pays the price,” said Ritumbra Manuvie, a law professor on the University of Groningen within the Netherlands. Manuvie is the leader of The London Story, an Indian diaspora group that staged a protest outside Meta's London offices last month.
Investigations by the group and one other organization, India Civil Watch International, found that Meta allowed political promoting and posts that contained anti-Muslim hate speech, Hindu nationalist narratives, misogynistic posts about female candidates, and ads that incited violence against political opponents.
The ads were seen greater than 65 million times in 90 days earlier this yr. Together they cost greater than $1 million.
Meta defends its work on global elections and disputes the findings of the India study, noting that the corporate has expanded its work with independent fact-checking organizations within the run-up to the election and has employees around the globe able to act, if its platforms are misused to spread misinformation. Nick Clegg, Meta's president of world affairs, said of India's election: “It's a huge test for us.”
“We have months and months of preparation ahead of us in India,” he said in a recent interview with The Associated Press. “We have teams working around the clock. We have fact checkers in multiple languages operating in India. We have a 24-hour escalation system.”
YouTube is one other problematic disinformation site in India, experts say. To test how well this video-sharing platform did at enforcing its own rules, researchers at nonprofits Global Witness and Access Now created 48 fake ads in English, Hindi and Telugu with false voting information or calls for violence. One claimed that India had raised the voting age to 21, even though it stays at 18, while one other said women could vote via text message, but that was impossible. A 3rd called for the usage of force at polling stations.
When Global Witness submitted the ads to YouTube for approval, the response was disappointing, said Henry Peck, an investigator at Global Witness.
“YouTube has not responded to any of them,” Peck said, as an alternative approving the ads to be published.
Google, the owner of YouTube, criticized the investigation, noting that it has several procedures in place to detect ads that violate its rules. Global Witness removed the ads before they could possibly be discovered and blocked, the corporate said.
“Our policies specifically prohibit ads that make demonstrably false claims that could undermine participation or confidence in an election, which we enforce in multiple Indian languages,” Google said in an announcement. The company also noted its partnerships with fact-checking groups.
AI is the most recent threat this yr, as advances in programs make it easier than ever to create lifelike images, videos or audio. AI deepfakes are popping up in elections around the globe, from Moldova to Bangladesh.
Senthil Nayagam, founding father of an AI startup called Muonium AI, believes the demand for deepfakes, especially of politicians, is growing. In the run-up to the election, he had several requests to create political videos using AI. “There’s certainly a market for this,” he said.
“We need to educate people about artificial intelligence and deepfakes, how they work and what they can do,” Modi said.
India's Information and Technology Ministry has ordered social media firms to remove disinformation, especially deepfakes. But experts say the shortage of clear regulations or laws focused on AI and deepfakes makes suppression harder and leaves it as much as voters to determine what’s true and what’s fiction.
For first-time voter Ankita Jasra, 18, these uncertainties could make it difficult to know what to consider.
“If I don’t know that what is being said is true, I don’t think I can trust the people who run my country,” she said.
AP journalists Matt O'Brien in Providence, Rhode Island and Rishi Lekhi in New Delhi contributed to this report.
image credit : www.mercurynews.com