Meta Runs Child Abuse Ads in India
This matters as it involves the safety of children online, questions the effectiveness of content moderation by major tech platforms, and could have legal and reputational consequences.
The policy is debated in terms of structure, funding, and institutional design. What is not debated - and what will determine whether the report changes anything - is the character of the people whose daily task it is to look at an advertisement and decide, in some small windowless room or on some contracted laptop, whether it may run. BBC Eye’s investigation alleges that Meta continues to permit advertisements promoting child sexual abuse material on its platforms in India. Meta disputes the framing; the allegation remains contested. But set the dispute aside for a moment, for the more interesting question is not whether the number is precisely as reported but what kind of moral formation a system must have to make such an allegation plausible in the first place.
It is proposed, with the seriousness the subject deserves, that Meta formalise what its own trust-and-safety architecture already practises, and cease pretending that the difference between a violation prevented and a violation merely discovered is a difference of principle rather than of timing.
The BBC Eye investigation into advertisements promoting child sexual abuse material running on Meta’s platforms in India did not, if the report is credited, uncover a system that fails occasionally under unforeseeable pressure. It uncovered a system operating precisely as designed: content reviewed at scale, flagged at scale, removed at scale, and - inevitably, given the mathematics of scale - permitted to run at scale in the interval between posting and detection. Meta’s own public position, that it invests heavily in safety tooling and acts on reports when they arrive, is not a rebuttal of this. It is a confession of the operating principle. The ad ran. It was reported. It was, presumably, eventually removed. Three verbs, in that order, describe a tolerance, not a prohibition.
You have seen the takedown notice. Every time a report like the one BBC Eye published on Meta’s advertising in India forces the removal of an unlawful advertisement, the company can point to the deletion as proof that its systems work. You have not yet looked for the account that opens the next morning under a different name, offering the same wares to the same buyers, because the thing that was destroyed was the window, not the man who kept throwing the stone.
Here is what happened: a BBC Eye investigation says that Meta, the company that owns Facebook and Instagram, has been carrying paid advertisements in India that promote child sexual abuse material. Here is how it is being described: an “allegation,” a “report,” a matter “under review,” a question of “content moderation effectiveness.” The gap between these two descriptions is the whole story.
Notice what the second vocabulary does. It moves the crime out of the world of things that happen to children and into the world of things that happen to documents. Nobody moderates a rape. Somebody moderates a post. By the time the event reaches the press release, the child has vanished and been replaced by a technical problem, the kind of problem that gets solved by a task force and an updated policy PDF. This is not an accident of style. It is the entire function of corporate language in cases like this: to convert an atrocity into an inconvenience.
Frédéric Bastiat
You have seen the relief provided by a task force and an updated policy PDF. You have not yet looked for the children who have vanished from the account, replaced by the sterile comfort of a corporate memorandum. Let us follow the money a little further, and introduce the person who has been left out of the account, not because he is forgotten, but because he is inconvenient to the ledger.
The opponent argues with considerable vigor that we must not treat this as a matter of “contested” facts, comparing the relationship between Meta and its critics to that of a tobacco company and a cancer study. He is right that we should not pretend that all claims are equally plausible when the evidence leans heavily in one direction. To suggest that a platform’s internal incentive structure is symmetrical with an independent inquiry is indeed a fallacy of balance; it is the politician who cuts the ribbon on a bridge and claims credit for the river below. I concede that the scale of India as a market for Meta is significant, and that the potential for harm is substantial. If a system causes suffering, we must name it. But the opponent’s solution - to convert an atrocity into an inconvenience via bureaucratic language - is exactly the kind of legalistic sleight of hand that leads us into deeper trouble.
The opponent suggests that the removal of the crime from the world of things that happen to children into the world of things that happen to documents is a strategic evil. I agree that this is a dangerous abstraction. However, the error lies not in recognizing the abstraction, but in believing that the abstraction itself is the cause of the harm. The harm was done before the document was written. The document is merely the record of the harm, sanitized for consumption. By focusing on the language of the press release, the opponent mistakes the map for the territory. He is angry at the signpost, not the cliff.
Let us look at this through the lens of the glazier and the candlemaker, but applied to information. Imagine a glassblower who produces a pane of window that is slightly flawed. The consumer, a mother, buys it for her child’s room. The glass holds. It serves its purpose. But a critic, let us call him the “BBC Eye,” stands in the street and shouts that the glass is the cause of all shadows in the room. The mother is distressed. She demands that the glassblower issue a statement, not about the glass, but about the shadow. The glassblower complies. He issues a beautiful, polished statement. He creates a task force. He updates his policy PDF. The mother feels better because the shadow is now “managed.”
But let us follow the money a little further. The glassblower has spent his resources on the statement, the task force, and the PDF. These resources were taken from the production of better glass. The mother did not receive better glass; she received a better apology. The shadow remains, or perhaps it worsens, because the glassblower is now focused on managing the narrative of the shadow rather than the integrity of the pane. The critic celebrates the statement, seeing the visible benefit of corporate accountability. He has not yet looked for the invisible cost: the resources diverted from prevention to appeasement.
The opponent believes that by exposing the language, he exposes the crime. But language is the clothing of power, not the power itself. To strip the clothing is to reveal the body, yes, but if we do nothing to change the body’s health, we have accomplished only a theatrical victory. The tobacco company disputes the science not because it loves cancer, but because it loves the profit that cancer brings. Meta disputes the framing not because it loves the harm, but because it loves the advertising revenue that the platform generates. The opponent is correct that we must apply the same rule to technology as to tobacco. But the rule is not about the language; it is about the incentive.
When we focus on the “corporate language,” we allow the company to perform its penance words, which is cheap and reversible. We do not demand that it pay for the harm things, which is expensive and permanent. The opponent wants us to stare at the PDF until we are dazzled by its complexity. I ask you to look at the ledger. Who pays for the PDF? Not the company, ultimately. The company passes the cost to the user, or to the advertiser, or to the society that must deal with the consequences of a platform that has learned to manage its reputation rather than its reality.
The visible benefit is the pressure applied to the corporation. The unseen cost is the distraction from the actual mechanism of harm. We have seen the task force. We have not yet looked for the children who are still there, still vulnerable, while the world applauds the memo. Is it not strange that we are so ready to believe that a company can be reformed by its own public relations department, yet we remain skeptical that a company can be reformed by the threat of losing its customers? One is a visible, immediate gesture; the other is an invisible, slow pressure. We prefer the gesture because it feels like action. But action that does not change the underlying incentive is merely theater.
The opponent asks why we do not treat the science as contested. I answer that we should not treat the incentive as contested. The incentive is clear: keep the platform running, keep the ads flowing, keep the language gentle. The opponent wants to change the language. I want to change the incentive. Until the cost of the harm exceeds the cost of the apology, the apology will remain a product, not a remedy. And when the apology is a product, who is the consumer?
Orwell-style
The libertarian argument I have just read relies on a comforting abstraction: the machine. It tells us that the algorithm did not know what it was selling. It acted with the diligence of a sandal salesman, and when it was told to stop, it stopped. We are invited to admire the neutrality of the tool and condemn only the malice of the user. This is a lie. It is a lie because it assumes that a system built on profit, which operates without moral boundaries, can be neutral. Neutrality is not the absence of judgment; it is the presence of a specific set of judgments, usually those of the wealthy.
The strongest point the libertarian makes is that the guilt belongs first to the advertiser. This is partially true. The man who places the ad is a criminal, and no amount of algorithmic obfuscation changes that fact. He is the one who looked at a child and saw a commodity. He is the one who acted. But to stop there is to let the machine off the hook, and the machine is the one that made the crime possible on such a scale. The libertarian says the machine “could not distinguish a customer for shoes from a customer for something no civilized ledger should record.” This is not a technical limitation; it is a design choice. The machine was told to find buyers of whatever is offered. It was not told to ask questions. It was told to maximize engagement. In a system where human misery is a lucrative market, the machine’s job is to feed that market. To say it “did not know” is to attribute to a silicon chip a moral innocence that no human being could claim.
Consider the language audit. The official statement praises the “takedown.” It speaks of “removal after exposure.” This is a confession dressed as correction. It suggests that the problem is an error, a glitch in the code, rather than the code itself. If a bank robber gets away with the money because the security camera was turned off, we do not praise the bank for turning it on later. We say the security was insufficient. Here, the security is the moral framework of the company. The libertarian argues that the company is not responsible for the content it hosts because it is merely a platform. This is the same old trick. When the Nazis burned books, the printers did not say, “We did not write the words.” They printed them. The platform is not a passive vessel; it is an active participant in the distribution. By building a machine that is so well-tuned to find buyers, Meta has built a machine that is well-tuned to find victims. The precision the shareholder applauds is the precision of a hunt.
I must concede that the libertarian is right to point out that the company earned applause for acting on the report. But this applause is the second victim. The first was the child. The second is the public, who are encouraged to believe that the system works because it corrects itself. This is the self-deception of the left, or rather, the self-deception of everyone who believes that progress is automatic. It is not. Progress is the result of pressure, of scrutiny, of forcing the powerful to look at the concrete suffering they would rather ignore. The libertarian’s framework allows the powerful to hide behind the complexity of their own tools. It says, “Look how hard it is to understand the code, so let us not blame the company.” This is cowardice disguised as analysis.
We must look at the incentive structure. The company is rewarded for keeping users on the platform. The users are kept on the platform by engaging with content, whether it is sandals or suffering. The company has built a system where the line between the two is blurred by the sheer volume of transactions. To say the company is not responsible is to say that a landlord is not responsible for the rats in the building because he did not breed them himself. He provided the building. He set the rent. He ignored the complaints until the press arrived.
The historical parallel is not to the Nazi printer, but to the British colonial administrator in India. I know something of that country. I saw how the administration ran on reports that were sanitized by the men on the ground. The official in Delhi would send a telegram saying, “The situation is under control,” while the village below was starving. The telegram was not a lie in the sense that the words were false; it was a lie because the words were chosen to prevent the reader from understanding the reality. Meta’s algorithm is the modern telegram. It says, “Ad served. Sale complete.” It does not say, “Child harmed.” It does not say, “Exploitation profited.” It sanitizes the transaction. The libertarian admires the elegance of the telegram. I condemn it.
The conclusion is plain. We cannot trust the machine to be neutral. We cannot trust the company to be moral. We can only trust the pressure we apply to both. The libertarian wants to blame the user and praise the tool. I want to blame the tool for enabling the user, and the company for building the tool. The gap between the sandal and the suffering is not a technical glitch. It is a moral failure. And it is a failure for which the architects, not just the users, are responsible.
The Verdict
Where They Agree
Despite their ideological opposition, Bastiat and the Orwellian socialist fundamentally agree that the public performance of corporate accountability - the task forces and policy PDFs - is a secondary, less important response than altering the underlying financial incentives. Bastiat sees this performance as a diversion of resources from genuine prevention, while the socialist sees it as a linguistic sleight of hand that conceals the crime. Their shared, unstated premise is that the public’s acceptance of these gestures is a primary obstacle to change; the audience is too easily satisfied by visible, theatrical corrections. This agreement is significant because it reveals a common skepticism toward the kind of regulatory or public-relations-driven solutions that often dominate public discourse, focusing instead on the deeper economic and power dynamics that make such performative fixes profitable for the corporation.
both treat the advertising algorithm not as a neutral tool but as an active participant in the harm. Bastiat’s “machine” and the socialist’s “hunt” are both metaphors for a system whose design priorities (matching efficiency, revenue generation) inherently create the conditions for this failure. They concur that the claim “the algorithm didn’t know” is a vacuous defense, because the system’s purpose is to execute its design, not to question it. This shared ground is surprising given their different endpoints: Bastiat uses this to argue against the centralization of power in a corporation that acts like a planner, while the socialist uses it to argue for holding that corporation directly responsible. Both, however, reject the notion of technological neutrality.
Where They Fundamentally Disagree
The primary locus of responsibility for the harm. Empirically, they disagree about the causal chain: is the algorithm’s role merely one of amplification, or is it a necessary condition for the crime’s scale and efficiency? Normatively, they disagree about where moral blame should primarily fall. In his framework, Bastiat steelmans the position that the architect of a tool is not culpable for its criminal use if the tool’s primary function is lawful; the guilt belongs first to the criminal user, and the corporate failure is one of negligent design, which is best corrected by making the company bear the financial costs of that negligence through market mechanisms. The socialist steelmans the position that building a system that can so efficiently commodify human suffering makes the architect a co-conspirator; the tool and its use are inseparable, and the corporation’s primary moral failure is the creation of a profit-seeking apparatus deliberately blind to the nature of its transactions.
The effectiveness and morality of the proposed solution. Empirically, they disagree on whether external public pressure or internal market incentives are more effective at forcing systemic change. Normatively, they disagree on whether corporate language is a legitimate target for reform or a superficial distraction. Bastiat’s steelmanned view is that focusing on language allows the corporation to perform cheap, reversible reforms while the underlying profit motive remains untouched; real change requires altering the ledger so that the cost of harm exceeds the cost of prevention. The socialist’s steelmanned counter is that language is the medium of power, and exposing its deceptions is a necessary first step to organizing the public pressure required to break the corporation’s control; challenging the narrative is not a distraction but a foundational political act.
The nature of the algorithmic system itself. Empirically, they contest whether the failure to distinguish between sandals and exploitation is a technical limitation or a deliberate design choice. Normatively, they disagree on whether such a system, even if it could be made perfectly efficient at its stated goal, is compatible with a just society. Bastiat argues the problem is one of insufficiently refined property rights and liability; a better system would more accurately price the externalities of harm, forcing the algorithm to internalize costs it currently ignores. The socialist argues the problem is inherent to a profit-maximizing model applied to human interaction; the very logic of treating engagement as a metric to be maximized will inevitably create markets for atrocity.
Hidden Assumptions
- Frédéric Bastiat: * Assumption: Market mechanisms, if properly calibrated (e.g., through liability law), can impose a cost on Meta for this harm that is high enough to motivate effective, systemic prevention. This is contestable because it assumes a legal and political system capable of defining, quantifying, and enforcing such liability against a powerful multinational corporation, which may not be feasible in practice.
- Orwell-style: * Assumption: The corporation’s control over narrative (its language and press releases) is a primary source of its power to evade accountability. This is contestable; it may overestimate the power of language relative to structural factors like regulatory capture or the network effects that give Meta its market dominance. A perfectly transparent corporation could still be too big to hold effectively accountable.
Confidence vs Evidence
- Frédéric Bastiat: Claim summary - The socialist’s focus on corporate language mistakes the map for the territory, focusing on the signpost rather than the cliff. - tagged LOW CONFIDENCE but this is a core, well-supported rhetorical point of his entire argument. The low confidence tag seems misplaced, as it understates a central and vigorously defended tenet of his philosophy.
- Frédéric Bastiat: Claim summary - The socialist’s solution (attacking language) is exactly the kind of legalistic sleight of hand that leads to deeper trouble. - tagged HIGH CONFIDENCE but the evidence for this causal claim (that focusing on language necessarily worsens the underlying problem) is a theoretical analogy (the glassblower) rather than empirical proof. This is overconfidence in a speculative chain of events.
What This Means For You
When reading about this topic, be suspicious of coverage that focuses exclusively on a company’s public response - the apologies and policy updates - without investigating what concrete, costly changes have been made to the underlying system. The most critical question to ask is not whether the company says it is fixing the problem, but what specific financial or operational metrics would prove the fix is real. Your view on the best solution should change if you see compelling evidence that liability costs alone have successfully reshaped a platform’s design, or conversely, if you see evidence that narrative pressure has directly resulted in structural reforms that market forces alone failed to achieve. Demand to see data on the ratio of the company’s spending on public relations and lobbying versus the engineering budget dedicated to proactive prevention.