The Surveillance Business Model

The surveillance business model is the economic structure that Meredith Whittaker identifies as the origin, not merely the context, of contemporary artificial intelligence. The model works by giving away services—search, email, social networking, maps—in exchange for behavioral data, which is then used to target advertising with increasing precision. Over two decades, the firms that mastered this model—Google, Meta, and their kin—accumulated three things in quantities no one else could match: data about human behavior at planetary scale, computational infrastructure to process that data, and the capital that advertising revenue generated. These three resources—data, compute, and capital—are precisely the inputs that large-scale machine learning requires. When the techniques of deep learning matured in the early 2010s, the companies positioned to exploit them were not new entrants with clever algorithms. They were the incumbents who already possessed the surveillance infrastructure. AI, in Whittaker's account, did not disrupt the surveillance economy. It was produced