In 2012 a cache of emails between members of a Russian pro-Putin youth group, Nashi, was leaked online. The emails revealed that young Russians were being paid to create thousands upon thousands of fake online personas, which they would use to douse any online news article that criticised Vladimir Putin with negative comments. YouTube videos from activists such as Alexei Navalny would be stifled by downvotes, while others were paid to generate blogs and videos to create the illusion of widespread support for the Putin regime.
Russia was far from the only government to use such practices. In 2013 a report by the think tank Freedom House recorded “paid progovernment commentators” at work in 22 countries, “particularly surrounding politically sensitive events such as elections or major street protests”. Russia stood apart, however, in its aggressive willingness to direct this new power across its borders. When Mark Zuckerberg was called to appear before senators in April 2018, he agreed that Russia had “repeatedly used complex networks of inauthentic accounts to deceive and manipulate people in the US, Europe and Russia”.
At the time, fake accounts were seen as a bad thing – Zuckerberg told the senators his company used “machine learning and artificial intelligence” to “block millions of fake account attempts each day”. But in 2025, attempts to maintain the internet’s hold on reality seem quaint. Facebook no longer appears to be interested in whether an online persona is backed by a real person.
On 27 December, Meta’s vice-president of product for generative AI, Connor Hayes, told the Financial Times that a new set of AI products would allow users to create AI bots that “exist on our platforms… in the same way that accounts do”. The bots, he explained, would “have bios and profile pictures and be able to generate and share content”. Before long, it is probable that much of what is posted on Facebook and Instagram (and perhaps to a certain extent WhatsApp) will not have been written by a human being. Bots talking to bots, sharing AI glurge, and upvoting is “where we see all of this going”, Hayes told the FT.
Back in 2018, as people were discovering that the people they interacted with online weren’t really as American or European as they claimed, it was also dawning on the public that they weren’t really the customers of the social media platforms they used. An aphorism of the media industry – “if you’re not paying, you are the product” – was widely repeated as social media users understood that they were providing something at their own expense that was sold on to advertisers.
On a rational planet this might have led a large majority of social media users to withhold their data and creativity from Zuckerberg until he agreed to pay them for it, but they didn’t. This emboldened technology companies to help themselves in an ever-more blatant fashion, siphoning off the world’s intellectual property into the large language models from which they now expect to draw their future revenue. Our political class has done less than nothing to prevent this; the Starmer government’s proposal is that businesses should have to “opt out” of their intellectual property being stolen to further enrich the world’s wealthiest people. No one should have to opt out of being burgled.
Who is all this for? Within a couple of years, AI infrastructure is expected to use more energy than the Netherlands and more clean water than Portugal. What are these huge resources being spent on? If the answer is advertising, it’s hard to see what an internet of bots offers to brands. In the long run, how many companies will pay to reach an audience of software that has no money or interests of its own? How many users will be happy for their pet “AI” to hawk cat litter or crypto on their behalf? Perhaps the answer is that Meta believes the internet can be made even more unnaturally compelling, that everything on it can be made even more obviously fake, but that consumers will be kept hanging around in an increasingly weird and coercive relationship.
Or perhaps the advertisers are the product too, and real product is not a commercial advantage but an idea – a narrative of profound economic and social change, driven by AI, that is being sold on financial markets. That idea is currently flying off the shelves: a single company, Nvidia, gained $2trn in market value over 2024. Mark Zuckerberg is about $80bn richer than he was this time last year, having pivoted his company from a big bet on virtual reality and the “metaverse” to a commitment to AI. It remains to be seen if that idea will become real, or if it too – like almost everything on social media – is a fiction.
[See also: Is BlueSky having its breakthrough moment?]
Content from our partners
Related
