An Indian-Australian IT engineer has spent more than a year documenting a surge of anti-Indian hate content spreading across Facebook and Instagram, finding that even minor interactions with mildly biased posts can rapidly funnel users — including teenagers — into an escalating spiral of racist and extremist material.
Mohit Sharma, who works as an IT engineer at an Australian high school, began tracking the problem after a student at his school was exposed to a racist post online. Concerned that toxic content was breaking out of the fringes of social media and reaching young people, he decided to conduct a systematic experiment — and what he found over the following 16 months he described as "horrible".
The Experiment: How Quickly the Algorithm Turns Hateful
Sharma created three separate accounts on Facebook and Instagram to observe how quickly a new user could be drawn into racist content. He tracked the algorithmic behaviour across both Meta-owned platforms, documenting what happened when accounts interacted with posts that contained even low-level negative depictions of Indians.
"The moment you interact a little bit, you suddenly fall into an algorithm full of, just, extreme hatred, misinformation and disinformation and AI-generated content," Sharma said.
He noted that the process was self-reinforcing: once a user's existing biases were confirmed by the algorithm, the content became progressively more extreme. Some accounts encountered AI-generated videos with captions calling for violence against India, while others were directed toward white nationalist messaging designed to inflame anti-immigration sentiment.
"Even if you have a little bias and you fall into that algorithm, those biases will be confirmed and then you'll become more hateful against one community," Sharma said. "That was horrible because if I'm a child, for example in a high school, how would that affect me in the future?"
He also found that posts with positive or sympathetic portrayals of Indians — including content about family tragedies — were being drawn into the same racist pipeline and targeted with abuse.
Scale of the Problem: Hundreds of Thousands of Likes
The anti-Indian posts Sharma documented were not fringe content attracting minimal attention. Anonymous accounts on Facebook and Instagram were accumulating hundreds of thousands of likes on abusive posts, spreading misinformation and harmful stereotypes under the guise of internet meme culture. The sheer volume of engagement suggests algorithmic amplification is playing a significant role in the content's reach.
Sharma compiled more than a year's worth of findings — covering what he described as "racist language, dehumanising statements, threats and descriptions of discriminatory conduct" — into a formal report released in late June. The report concluded that the evidence revealed "recurring patterns of hostility, amplification, recommendation behaviours and political messaging that warrant further independent examination."
It warned that, if left unaddressed, such online environments "may contribute to the normalisation of prejudice, increased social division, declining trust between communities."
Meta's Response Falls Short, Community Feels the Strain
Sharma says that attempts by users to report the hateful content to Meta have largely been ineffective. Where offensive posts were previously removed promptly, he says that is no longer the case.
"It's happening too much and a lot of people try to complain about those posts … and even after reporting, they aren't disappearing," he said. "In the past, they used to be taken down immediately but now it's not happening."
The impact on the Indian-Australian community is tangible, according to Sharma. He says people within the community are internalising the hostility and beginning to feel like outsiders — a sentiment he fears could persist for generations.
"A lot of young people are extremely stressed, exhausted and think it's bound to happen and normal," he said. "It's not good."
The findings arrive amid a broader global conversation about Meta platform reliability and accountability, as well as growing scrutiny of how social media algorithms recommend increasingly extreme content to users. Sharma's report calls for independent examination of the patterns he documented, suggesting the issue extends well beyond individual posts and into the structural design of the platforms themselves.
