Zuckerberg and AI therapists: Watch your minds!

Statements made by Meta CEO Mark Zuckerberg regarding future virtual relationships with artificial intelligence (AI) companions and AI therapists are currently a hot topic.

First, his comments on Dwarkesh Patel’s podcast drew attention. Zuckerberg was discussing the future of AI friends, therapists, and girlfriends.

According to Meta’s founder, while Americans, on average, have only three friends, they “wanted fifteen friends.” He then argued that although emotional bonds with AI bots are not currently socially accepted, society would eventually “find the words” to understand that people using AI to fill the loneliness and void in their lives are “rational.”

Zuckerberg continued to touch upon this subject. The Meta CEO’s less-noticed remarks, made a few days prior on Ben Thompson’s “Stratechery” podcast, further elaborate on his vision of how AI companionship might function.

Many interpreted Zuckerberg’s words to mean that you would have AI friends instead of real friends, and in fact, that’s more or less what he meant:

“There’s an interesting sociological finding: the average American has fewer than three friends, and the average American wants to have more than three friends. So, ideally, you want to enable people to connect with the right people, and that’s something we try to help people with. When they’re not physically together, they can stay connected through our apps, keep in touch with people, meet new people. But going forward, I think there’s going to be a dynamic where you’re interacting with different people on different topics.”

However, there’s something more significant (and ominous) that the tech billionaire implied between the lines: the fact that Meta has an AI strategy built on knowing much more about your friends and family.

In his interview with Thompson, Zuckerberg stated:

“I think one of the things that I’m most focused on is how AI can help you be a better friend to your friends. There are so many things that I don’t remember about people I care about, that I could be more thoughtful. There are issues like, I’m a ‘plan at the last minute’ kind of person, and then issues arise like, ‘I don’t know who’s around, and I don’t want to bother people.’ An AI that has good context on what’s going on with the people you care about can help you with that.

Good personalized AI isn’t just about having some basic information about your interests; a good assistant or good personalization is about having a theory of mind about how you think about things. So, this is what we do with all of our friends. It’s not just like, ‘Okay, this is my friend Bob, and he likes this thing.’ You deeply understand what’s going on in that person’s life, what your friends are going through, what their challenges are, and what the interplay is between these different things.”

Elsewhere in the interview, Zuckerberg points out how interaction provided on Facebook has changed with tools like Instagram. “It used to be that you would interact with the people you were connected to in the feed,” explains the Meta CEO, “for example, someone would share something, and you would comment, and that’s how your interaction would happen.”

So, what’s the situation now? Zuckerberg explains clearly:

“Today, we see Facebook, Instagram, Threads, and I guess now the Meta AI app, and many other things we do, as discovery engines. Most of the interaction doesn’t happen in the feed. The app works like a discovery engine algorithm to show you interesting things, and then the real social interaction happens when you find something interesting and add it to a group chat with your friends or a one-on-one chat. So, there’s a flywheel effect between messaging, where the real, deep, and nuanced social interaction happens, and the feed apps, which are increasingly just becoming discovery engines.”

The Meta CEO doesn’t hide that they are designing this as a “business model.” This model perhaps represents the pinnacle of subjecting both the worker and society as a whole to the “logic of capital”:

“[We] want to use AI to basically enable any business that wants to achieve a certain business outcome to come to us and get service without needing to produce any content or have any information about their customers. They should just be able to say, ‘This is the business outcome I want, this is the fee I’m willing to pay, I’ll connect you to my bank account, I’ll pay you for the business results you achieve’… I think this is a redefinition of the advertising category. If you think about what percentage of GDP advertising is today, I would expect that percentage to increase.”

This is a rather critical statement. Zuckerberg is essentially saying: Businesses will not have to produce any content or know anything about their customers. Meta, or rather Meta’s AI bot, will take over the connection between businesses and customers and most decisions related to branding. It will have more data, a larger scale, more connections, and the world’s largest black box. In the future, marketing and advertising for all companies will mean delegating commerce to an automated infrastructure controlled by a single person (or bot).

What better “social engineering” could there be?

This “business model” also points to a future that will eliminate the “public-private distinction,” one of the hallmarks of bourgeois civilization. Zuckerberg mentioned back in 2010 that he wasn’t hiding his vision of such a “humanity”:

“The days of you having a different image for your work friends or co-workers and for the other people you know are probably coming to an end very quickly. Having two identities for yourself is an example of a lack of integrity… It’s a big challenge to get people to a point where they can be more open. But I think we’ll get there.”

Let me remind you that Zuckerberg has taken quite a few steps in this regard. For example, in 2007, he launched Beacon, which automatically added your Facebook purchases to your feed. This application exposed users’ HIV statuses and which engagement rings they bought.

Moreover, recently, the Wall Street Journal published a story: Meta’s chatbots were talking about fantasy sex with children.

Meta allows “synthetic personalities” to offer full-scale social interaction, including bantering via text, sharing selfies, and even engaging in live voice chats with users.

What is happening once again confirms one of Marx’s analyses regarding the behavior of capital. In Capital, Marx distinguishes between “formal” and “real” forms of subsumption. Initially, capital absorbs the existing labor process—that is, the techniques, markets, means of production, and workers—into itself. Marx calls this “formal” subsumption.

In this process, the entire labor process continues as before, but the capitalist, who monopolizes the means of production and thus the workers’ means of subsistence, forces the worker to submit to wage labor and can accumulate capital using existing markets.

However, capitalism cannot develop on the limited foundations of existing productive forces. The preconditions for the actual capitalist labor process can only be created by capital itself. Thus, capital gradually transforms social relations and forms of labor until they are completely intertwined with the nature and requirements of capital, and the labor process becomes truly, really subsumed under capital.

Therefore, for capital to accumulate, to ensure that property owners do not become propertyless, it must develop models and labor processes that subject not only wage labor but all of society to itself.

Your relationships with friends, what you experience with your family, even information about your mental health, must therefore be laid out before capital:

“Personally, I believe everyone should have a therapist. A therapist is like someone they can talk to throughout the day, or if not throughout the day, about whatever they are worried about. For people who don’t have a therapist, I think everyone will have an AI assistant.”