This should be fine, pose no problems and create a sustainable pathway to future prosperity for us all. Right?
Over the weekend, the Wall Street Journal reported that Meta CEO Mark Zuckerberg is so enamored by the possibilities of artificial intelligence that he hopes to build AI agents that will undertake everyone’s work, and he’s using himself as a guinea pig.
As per WSJ: “Zuckerberg […] is building a CEO agent to help him do his job, according to a person familiar with the project. The agent, which is still in development, is currently helping Zuckerberg get information faster — for instance, by retrieving answers for him that he would typically have to go through layers of people to get, the person familiar with the project said.”
Conceptually, Zuckerberg is hoping that he can upload all of the information about his daily activity into an AI system, which will eventually learn enough about his actions and responses to simulate his professional self.
Which could work, in theory. Current AI systems are built on pattern recognition, and they map out likely responses based on keyword matching against a whole corpus of response data built into their back end. If that system were trained solely on one person, over enough time, it could emulate their likely responses and estimate the path they’d be likely to choose, based on their past actions.
I mean, that says nothing for free will, and people’s capacity to learn new information and evolve our approach over time. But if someone wanted an AI agent to learn their processes and replicate them with AI decision-making, it seems plausible given the current systems.
Of course, Meta is also looking beyond the current limitations of AI as it seeks to develop “superintelligence” and digital systems that better replicate the human brain. That’s a more ambitious goal, which is well beyond the data connections that power current AI systems. But maybe this project could help to bridge the gap between where we are now and where Meta wants to be, with more personalized, individual decision mapping to help understand and replicate each person’s responses.
Meta has also discussed this in a social media context as the company builds AI tools that are designed to better understand each user’s preferences over time. Those tools may eventually be able to deliver more individually engaging experiences as a result.
So there is business value beyond just replicating a virtual self. As Meta continues to invest hundreds of billions of dollars into massive data processing centers to parse that data, it seems plausible that the company could eventually enable AI agents for millions of people, highly attuned to each users’ behavior.
So long as these AI agents don’t evolve too much.
Which does seem like a limitation. Though for Zuckerberg, building a personalized AI assistant has been a longer-term ambition, and one that has clearly stuck with him for many years.
Back in 2016, Zuckerberg showed off his self-created AI home assistant to CNN News. The presentation involved Zuckerberg interacting with it using voice commands.
Zuckerberg’s awkward exchanges and daily habits became more of a story at the time than the actual functionality of the tool (which was voiced by Morgan Freeman). However, it was an early indicator of Zuckerberg’s broader desire to create an AI-powered system that would be able to automate and assist him with elements of his daily life.
I’m not sure everyone will share the same enthusiasm, but clearly, Zuckerberg is hoping to drive expanded take-up of similar tools, in both professional and personal settings.
That’s what’s driving his enthusiasm for AI, and the expansion of Meta’s AI projects, which have now become the biggest project within the business.
Will that pay off? A lot depends on broader AI adoption, but it’s a strong indicator of where Zuckerberg is headed with all that money being poured into his AI projects.
Though if AI can replicate people professionally and personally, what happens to humanity at the end of that chain?
