Search engines are incorporating imitative AI in its search results at the same time as it is being inundated with garbage AI generated spam. Bing is doing something similar and driving AI into all its products, such as Windows and Office, via the Copilot products. Meta is forcing an AI chatbot into all it apps, including your personal news feeds and search. These initiatives are already making these products worse. It seems silly that these companies would deliberately worsen their primary consumer-facing products, but there we are. The question is why? And the answer might be, to a significant extent, the failure of antitrust.
Each of the companies in question are either monopolies or close to it. They all leverage their near-monopoly positions in one area to drive away competition in others. And the AI gold-rush is a good example of how that harms consumers and other businesses.
First, Google and Microsoft make a lot of money from their cloud services being used to run AI systems. The much-ballyhooed investment in OpenAI by Microsoft consisted significantly of cloud computing credits. Google, if it can make AI ubiquitous, stands to benefit in a similar fashion. So they use their power in search and/or operating systems and/or office applications to drive AI consumption. That in turn pushes other companies to do the same, which in turn drives spending on their cloud platforms.
Meta has no cloud platform, but by pushing AI it intends to create content that will drive engagement, and thus ad revenue. And, of course, Google benefits form every ad that is tied to AI spam.
These companies can do this because of the failures of antitrust. Facebook dominates social media in large part because it was allowed to buy up its competitors like Instagram. Google is in the middle of an antitrust trial showing how it has leveraged payola to protect its search monopoly and its search monopoly to drive other business, such as ads. There are few practical replacements for Microsoft’s office and operating system products. People are limited in their choices, and so these companies can actively damage their primary products in the pursuit of money in other business areas because they effectively have no competition in those areas.
For forty years we have allowed companies to consolidate, leverage power in one area to crush competition in others, and generally create a rentier class that is immune to business consequences. The ubiquity of AI that provides false information and disinformation is a consequence, in part, of this failure to police business. There are, of course, other contributing factors. But these companies clearly believe they won’t suffer consequences for damaging their consumer facing businesses in pursuit of AI gold. And they are probably correct, unless and until we start taking antitrust seriously. Companies should be terrified of the damage they are doing to their signature products. That they ar enot is an indictment of the way we have allowed capitalism to evolve.
If you want to solve the problems of AI — environmental degradation, disinformation, misinformation, etc. — then you need to start with strong antitrust enforcement. Without that, these companies have no reason to care about what AI does to their primary products.