The Algorithm's Perfect Corporate Colleague: Meta's Encounter with the Indian Government

Algorithms, unlike politicians, bureaucrats, or journalists, rarely hold press conferences. They do not explain themselves, issue clarifications, or blame the copy desk. They simply do what they are designed to do, until one day they do something spectacularly inconvenient. Then, suddenly, humans reappear to explain that the algorithm did not quite mean what it did. That, in many ways, sums up Meta’s latest encounter with the Indian government. The company apologised after Prime Minister Narendra Modi’s Facebook post briefly disappeared from the platform, attributing the episode to an AI-powered automated system. The explanation sounded familiar. In the age of artificial intelligence, the algorithm has become the perfect corporate colleague. It works round the clock, processes billions of pieces of content, and, every now and then, becomes the designated recipient of all blame. One can almost imagine the scene inside a technology company. A dashboard flashes red. Someone says, “Who removed the Prime Minister’s post?” There is an awkward silence before everyone turns to look at the algorithm. The algorithm, being a software, says nothing. The communications team, however, says plenty. The irony is that algorithms are no longer silent backroom tools. They decide what trends, what disappears, what goes viral, and, increasingly, what people believe everyone else is talking about. Social media companies often describe themselves as neutral intermediaries, mere highways on which information travels. Yet these highways appear to come with traffic lights, express lanes, and occasionally even VIP corridors. If a platform merely hosted content, there would be little debate. But today’s digital platforms do much more. They recommend videos, rank posts, personalise feeds, amplify trends, and, as governments across the world increasingly highlight, shape public discourse itself. The invisible editor is no longer sitting in a newsroom. It sits inside millions of lines of computer code. That is perhaps why the government’s point over intermediary status deserves more attention than the temporary disappearance of a single post. The real question is not whether an algorithm made a mistake. Software inevitably does. The larger question is whether companies can simultaneously claim enormous influence over what billions of users see while insisting they are merely passive hosts whenever accountability comes knocking. Technology companies, of course, are hardly alone in discovering the convenience of automation. Every sector now has its version of “the system”. Airlines blame computer glitches. Banks blame server downtime. E-commerce firms blame technical errors. Somewhere in every organisation there appears to be a mythical creature called “the system”, which takes difficult decisions without consulting anyone and politely disappears before the inquiry begins. Artificial intelligence has simply upgraded that excuse. To be fair, moderating billions of posts across hundreds of countries is no easy task. Perfection is impossible and mistakes are inevitable. The challenge is not whether errors occur but how platforms respond to them and whether the rules governing those decisions are sufficiently transparent. Ironically, AI is becoming better at creating content than explaining why it removed it. Ask a chatbot to write a sonnet and it probably can. Ask an algorithm why one post stayed while another vanished, and the answer often resembles a weather forecast: something happened somewhere because multiple factors were involved. The debate is unlikely to end with one meeting in Delhi or one apology from Silicon Valley. Governments everywhere are asking similar questions about algorithmic accountability, paid amplification, and online safety. Platforms, meanwhile, continue to maintain that innovation requires flexibility and that excessive regulation could stifle the digital economy. Perhaps the only undisputed winner in all this is the humble apology. It remains the smallest button in the technology ecosystem, pressed only after the much larger buttons labelled “Remove”, “Boost” and “Recommend” have already done their work. Maybe future versions of artificial intelligence will become so advanced that they will not only decide what deserves attention but will also draft the apology before anyone asks for it. Until then, humans will continue to explain the algorithm.