Instagram child abuse ads were served to an undercover BBC reporter's account within days, according to the 2026 documentary The Careless Machine. The film traces paid advertisements containing child sexual abuse material from Instagram feeds to Telegram bots, and records Meta's response to the reporter's complaint.
What the BBC Found in The Careless Machine
The BBC World Service documentary The Careless Machine, published in July 2026, found that Instagram served its undercover reporter paid advertisements containing child sexual abuse material, alongside explicit adult ads, within days of a new account following suggestive influencer content. The ads offered videos for sale and redirected to bots on Telegram, the messaging platform.
Reporter Divya, working from a clean phone with an alias account, followed a chain that started with ordinary influencer videos and escalated within roughly a week. Ads displayed text offering "child video," "rape video," and "force video," with some depicting children of around school age in sexual scenarios. She did not need to click any ad for the feed to keep serving them.
The film's core claims, and how they are evidenced, line up as follows:
| Finding | Evidence shown in the film | Who must answer |
|---|---|---|
| Ads with nudity and porn on a new account | Screen recordings of the reporter's feed | Meta, owner of Instagram |
| Paid ads containing child sexual abuse material | Ads linking to Telegram bots offering videos | Meta, plus Telegram and Indian police |
| A reported CSAM ad ruled non-violating | Instagram's written reply to the reporter's complaint | Meta's ad review systems |
| Platform identified as a top tip-line source | Interview with Hyderabad cyber police | Platform accountability under Indian law |
Meta told the BBC that it fights child exploitation on its platforms, removes ads that violate its policies, and reports such content to the authorities. That statement sits beside the film's central tension: the company's own detection systems were flagging and reporting child sexual abuse material elsewhere on the platform, yet the ads it was paid to promote kept appearing.
How Instagram's Algorithm Escalated a Normal Feed
The escalation followed a repeatable sequence: engagement with one suggestive video trained the recommender to serve progressively more explicit content, ending in paid ads within about a week. The reporter started by watching a benign influencer clip. As she kept watching, the algorithm recommended more influencers, then more explicit videos she had never searched for.
Three stages stand out in her account:
- Interest capture: the recommendation system registered what held her attention and pushed similar influencer content.
- Escalation: within days, the feed included outright pornographic ads, despite Instagram's published rule that ads may not contain nudity.
- Paid abuse material: ads featuring children in sexual scenarios appeared, some with explicit sales text, without any click from her.
Brian Boland, a former Meta vice president who spent 11 years at Facebook in advertising and product leadership, described the mechanism to the BBC. In his framing, the ad system draws from an endless pool of possible advertising and tests slightly more extreme, more tantalizing material to hold attention. It optimizes for engagement and revenue, and, as he put it, it is an absolutely careless machine. He testified against Meta over child safety in 2026, saying safety investment remained small and siloed even as evidence of feed-related harm grew.
The film also notes a regional asymmetry in his testimony: Boland said the global south and non-US, non-Europe regions receive what he called dramatic underinvestment in safety. That matters for an India-specific finding, because the reporter's account operated entirely within Indian-language content ecosystems.
Where the Ads Led: Telegram Bots and Priced Clips
Every ad the reporter clicked redirected to a Telegram bot that offered a menu of child sexual abuse videos for sale, with catalogues labeled in Hindi transliterated into English, including terms for mother-son, father-daughter, and brother-sister content. One ad showed a man with his arm around a girl and text claiming he was 52 and she was 12. Another showed a crying girl of about seven with sales copy describing her being hurt.
Hyderabad's cyber police, whose child protection unit claims more than 400 arrests last year for watching or sharing such material, told the BBC that clips of abuse sell on some platforms for as little as 30 rupees, under £1. Shikha Goel, who leads the unit, explained the economics directly: viewers are the customers, and demand is what funds the creation of the material.
Her unit receives most of its tip lines from Instagram and Facebook, though she cautioned that this may reflect stronger platform detection rather than platform concentration of abuse. Notably, she said her team had not seen ads featuring child sexual exploitation material reported to them before, which is itself evidence of a reporting gap between Meta's ad systems and Indian law enforcement.
The reporter reported the ads and Telegram groups to the Indian authorities, and filed a complaint with Meta about the ad showing the crying seven-year-old. Instagram's reply, shown in the film, said the ad did not violate its guidelines and could not be removed.
What Meta Says, and What Its Own Records Show
Meta's official position to the BBC is that it fights child exploitation, removes policy-violating ads, and reports offending content to the authorities. The film juxtaposes that statement with reporting on Meta's ad business more broadly.
Two earlier investigations provide context. Reporting by The Wall Street Journal documented ads for illegal substances running on Facebook and Instagram and steering users to online drug marketplaces; Meta said it works with law enforcement against such activity. Reuters reported in 2025, citing internal Meta documents, that the company earned roughly $16 billion from ads for scams and banned goods; Meta said the documents present a selective view and that it focuses on rooting out fraud.
The documentary's question is narrower than either of those stories: is Meta scanning and reporting the content it is paid to promote? The evidence in the film suggests the automated ad review pipeline, backed by increased reliance on technical solutions and selective human review, did not catch ads that a public complaint had flagged. Boland told the BBC that decisions about how strictly ads are monitored involve the most senior people, naming Adam Mosseri at Instagram and Mark Zuckerberg at Meta.
Zuckerberg's own words appear in archival footage describing a 2009-2010 cleanup of scammy Facebook ads, when he says he removed a massive part of company revenue for user safety. Boland's answer when asked whether that is doable now was blunt: not while Zuckerberg runs the company. That is a participant's characterization, and Meta disputes the framing, but it is the film's explanatory core.
Can Indian Law Hold Platforms Accountable?
Retired Supreme Court of India Justice Madan Lokur told the BBC that India's current safe harbour interpretation makes it hard to hold a platform criminally liable for user-uploaded content, but argued that paid advertisements change the picture. His reasoning: a platform cannot claim it merely hosts content when it accepted payment to distribute it and earned money from it.
Safe harbour, the legal principle protecting intermediaries from liability for third-party content, was designed for hosting, not for a paid promotion pipeline. Justice Lokur argued the courts and government should look beyond it. Two remedies are available under existing powers:
- The Supreme Court of India could take suo motu cognizance of the issue, acting without a filed petition.
- The government can issue blocking orders and takedown orders against any social media platform if it chooses to act.
The Supreme Court of India has not, on the evidence in this film, taken such action on these ads specifically. Justice Lokur's framing sets the current boundary: what the film documented is, in his words, serious enough for the court to act on, but the legal machinery has not yet moved. Readers should treat his view as an expert legal opinion within the film, not a ruling.
Who Answered: Meta India's Arun Srinivas
Arun Srinivas, managing director of Meta India and formerly the country's ads business director, declined to answer the BBC's questions when confronted publicly at an AI summit in Delhi. The reporter asked him directly who was responsible for Instagram serving pornography and child sexual abuse material as ads, noting his prior role over the ads business.
Earlier in the same event, Srinivas had presented Meta's AI advertising vision, describing automated targeting so precise that a pickle seller could reach the exact person craving pickle at that moment. The film's point in placing the two moments side by side is that the same targeting machinery celebrated for commerce was, on the evidence shown, selling abuse material. Srinivas did not respond on camera, and no Meta India executive gave the BBC an interview.
FAQ
- What is The Careless Machine? It is a BBC World Service documentary published in July 2026 investigating harmful and illegal advertising on Instagram in India, including ads containing child sexual abuse material and the Telegram bots they link to.
- Did Meta know about these Instagram child abuse ads? The film shows Meta's systems were detecting and reporting child sexual abuse elsewhere on the platform, and that the reporter's complaint about one ad received a reply saying it did not violate guidelines. Former Meta vice president Brian Boland testified in 2026 that safety investment remained small and siloed; Meta says it fights exploitation and removes violating ads.
- Is Telegram responsible for the material the ads linked to? The ads redirected to Telegram bots offering videos for sale. The reporter reported these bots to the Indian authorities alongside the ads. The film documents the funnel but does not examine Telegram's own enforcement response.
- Can viewers report this material? Yes. In India, reports go to cyber police units such as Hyderabad's child protection unit, which processes tip lines from platforms. Internationally, the NCMEC CyberTipline at report.cybertip.org accepts reports of child sexual exploitation in the United States.
- Has anyone been held legally accountable? Not on the evidence in the film. Retired Justice Madan Lokur explains that India's safe harbour doctrine currently shields platforms from liability for user content, though he argues paid ads justify going beyond it, and that the Supreme Court or government could act.
From Watching to Publishing: Why Written Records Matter
This investigation mattered because someone turned hours of recorded screen evidence, police interviews, and testimony into a structured story that regulators and courts can act on. Investigations like this one live inside a 27-minute film, which is powerful for a broadcast audience but hard to cite, quote, or search.
If you have valuable reporting, interviews, or explanations sitting in YouTube videos, Skala Blog can turn that footage into a written article: paste the video URL, transcribe it, and generate a structured piece that readers and search engines can actually find. For anyone documenting serious subjects, a durable written record is part of the accountability itself.
Fork this article
Start a new branch from the same video, shaped your way. You keep the credit; the original keeps the attribution.
A fork in another language is filed as a translation of this article, so the two pages point at each other. You can unlink it later from the editor.
0/240
You are creating
- Format
- For
- Language
- Source
- Your angle
You will be asked to sign in before it is generated.
Buy credits