Eric Posner on Cultural Monopsonies and Data-As-Labor

On cultural monopsonies and data-as-labor

The latest and most interesting manifestation is the market for data labor. Starting with Jaron Lanier, commentators have pointed out that Facebook, Google, and other big tech companies have exploited a cultural understanding that when one uses internet services as a consumer one doesn’t “work” by supplying data despite the great value of the data for the internet companies, which use it to improve their AI services. This is not very different from my other examples, especially the women’s work example—since women were, in a rough sense, paid in kind for their work though not give their marginal product as they would in a market. The big tech companies would get better data if they paid people with money for their specific contributions rather than in kind through the general service provision (and, in fact, sometimes they do, but in a way largely hidden from the public, as when they hire people (as “workers”) to label photos), but very likely do not do so because they do not want to disturb the cultural monopsony from which they benefit. The result is that a huge number of people who make significant contributions to the development of AI and other services are unpaid, except roughly in kind, which not only retards the growth of technology services, but results in a massive transfer of wealth from ordinary people to the tech titans.