Big Brother sees all but shows no evil. This might sound like a line from dystopic fiction, but it is also an apt description of life in the Age of Facebook. The latest victim of Facebook's 'show no evil' policy is Marianne, the French personification of liberty. Facebook recently banned an advertisement for using the painting, Liberty Leading the People, by Eugène Delacroix. No, it was not copyright laws that had Mark Zuckerberg and his team worried. It was a glimpse of Marianne's breasts. In the painting, liberty leads the revolution oblivious to — and unashamed of — her state of undress. It depicts a historic moment, but all Facebook, rather its artificial intelligence 'moderators', could see was that liberty fell into what the company classifies as nudity. Facebook has since apologized for its mistake.
But Facebook and fig leaves go back a long way. The girl in Nick Ut's famous photograph of a napalm attack in Vietnam, the 30,000-year-old Venus of Willendorf and The Little Mermaid sculptures are just some of the others who have fallen prey, at least at first, to Facebook's policy against nudity. This, in spite of the fact that photographs of artwork depicting nude figures are allowed on the platform. Herein lies the problem: an AI programme cannot see the context of an image or a post. Its excisions are made on the basis of markers and keywords that it is trained to recognize. This is dangerous on several levels. First, it allows those in power — the ones paying for the development and training of these AI bots — to exercise control over information that goes out into the public domain. The line between censorship and surveillance can indeed be thin. More worryingly, the powerful conveniently put the blame for any censorship onto mechanical glitches, making it seem innocuous and unintentional. Facebook's human editors, however, are as adept at wielding the scissor as their AI counterparts. A study shows that 55 per cent of the time, content reviewers make the wrong call on blocking hate speech, like banning a post against a Nazi rally simply because it came with an image of a man sporting a swastika. It is difficult to blame them. Not only do they have to review one post every 10 seconds but they are also low wage earners who cannot afford to risk the displeasure of their employer.
Facebook is not the only culprit. Other tech giants like Google and Twitter use similar methods to decide what is fit for public consumption. The problem lies in what informs these decisions. These platforms — the nucleus of news distribution — are run by technocrats not elected representatives. They care not about sentiments of the public but of shareholders and State agencies that control their businesses. Whether it is kowtowing to the overarching anti-hate speech law in Germany or the more ludicrous bans on Winnie the Pooh in China, these businessmen would always err on the side of caution. The internet was based on the gospel, "information wants to be free" — free from censorship and free of charge. But if the businessmen running the internet have to choose between pesky free speech advocates and furious politicians angry with online content, there is only going to be one outcome: liberty will be vanquished.
