Guess what the biggest danger to the future of AI is

Advertising looms as the worst threat to the development of business models based on generative AI assistants that could prove to be very interesting tools, but whose value proposition could be destroyed by all kinds of arbitrary biases.

An interesting article by Nathan Sanders and Bruce Schneier in MIT Tech Review, “Let’s not make the same mistakes with AI that we made with social media”, explores an idea that I have written about on many occasions: what destroyed the value proposition of social networks and turned them into the cesspits they are today was the introduction of hyper-targeted advertising as a business model.

Sanders and Schneier argue that hyper-targeted advertising created a huge bubble of wild expectations about its efficiency and the results it promised companies. Advertising was already the business model for social networks and the internet in general, working in the same way as traditional media had for decades: context-based targeting without the need to capture users’ personal data. When Google and, later, Facebook (now Meta) began to explore the idea of storing user data and building detailed profiles on it, they attributed spectacular efficiency to that type of advertising; and so was born the myth of the laser-focused sniper that could place an ad for a particular product or service in front of the…