The Risk of Total Knowledge in the 21st Century

In 1941, a short story by an Argentine writer named Jorge Luis Borges was published in a book titled The Garden of Forking Paths. The story would mesmerize linguists, mathematicians, philosophers, and eventually computer scientists. The name of the story was The Library of Babel, and its thesis and mathematical depth made it a profound read.

I came across the story almost a year ago and have been fascinated with the concepts presented in it. The story presents an infinite library that contains every book that has ever been written and every book that will ever be written. The librarians make this claim because the books include every possible ordering of 25 basic characters and are limited to 410-pages. The only problem is that the majority of the books are complete gibberish. The librarians search endlessly for texts that contain all useful information, including predictions of the future, biographies of any person, decrypted messages from army communications, a letter written to a lover in 1621 — everything exists within the library. The issue was that the search for anything meaningful takes infinite time. The librarians searched for a book of secrets that they couldn’t find because they didn’t have enough time.

Borges wrote this story before the computer age, but now that it is here, we are reaching a point where this limitation, while still mathematically valid, will cease to be a problem. Let me explain how. From my interpretation of the story, mankind itself is who applies meaning to the books within the library. The search of the librarians is a map from their lives’ experience to text written in one of the books. There was a limit to how much human experience could be converted into knowledge. That limit was the number of humans alive and the information and experiences that each individual had access to.

I spent the last three and a half years of my life working with artificial intelligence, and upon my reading this story, an epiphany hit me. Computationally it’s impossible to generate all permutations of books. And even if it were possible, sorting through the books after they had been generated would take an equally infinite time. What artificial intelligence, deployed on future computing hardware, could allow is the ability to synthesize human experience, develop experiential maps, and find useful information at record speeds.

Synthesizing human experience is something that is happening today on a grand scale. All around the world, data is mined like a gold rush. Storage is cheap, and our human experience is free for the taking because we don’t care — everything we do online to tracked, sold, and modeled. We say, “I have nothing to hide,” “I’m getting useful services from Google and Facebook,” and then shrug and say, “the government is doing it anyway.” What I don’t think we realize is that once there is enough data, enough experience, then governments and corporations do not need to compute the entire library. Instead, they will have the most robust map of intelligence that has even been achieved.

The map of meaning, mined from our experience, will allow any who hold such power to find those hidden books of the library to see what the future will hold. You can claim that the mining of experience will have to be continually explored by humans for the system to find more knowledge as time progresses — but that may be a false assumption. We already have situations in which self-driving cars are trained using simulators or even video games like Grand Theft Auto 5. Knowledge and experience can be trained into models, even from a simulation.

The more pertinent observation is that given a robust enough simulation, ran on a future supercomputer, and explored by the experiential maps trained into the artificial intelligence, total (useful) knowledge could happen. In fact, it could happen sooner than we think. Every data point we give, every shrug toward privacy that we make, brings us one step closer to this inevitability. The concern for me is who gets there first. Is it a corporation? Google? Amazon? China? The US? With total knowledge comes total power. What are the implications of that power? We need to start thinking about this now. We need to start fighting for our data and privacy. Because it is no longer “I have nothing to hide,” it becomes a matter of humanity’s freedom in the future.

I have faith in humanity, so much faith, but I have less faith in governments and corporations attaining total power over information. This is one of the biggest existential threats we will face this century.

If Borges were alive today and saw the data acquisition, artificial intelligence, and leaps in computational power, how would his story be different? The librarians who endlessly search the library only had their individual or communicated experiences as a map to traverse the infinite expanse of information. What if those librarians were machines and had a map of all humanity’s experiences?

I challenge you to think about this and read The Library of Babel. This short story from 1941 may be more relevant to our modern era than we think.