Maybe some of you have read “Dial F for Frankenstein?” In that short story Arthur C Clarke imagines the consequences of a minor technical improvement in the global communications network. Suddenly a new lifeform emerges out of the network and it quickly learns how to protect itself and control its body. Clarke’s story is often cited as an example of what it might be like if a mind were to spontaneously emerge on the Internet.
It is a fun read and a provocative idea but, I think that spontaneous sentience emerging from the Internet is unlikely. This is simply because most computers are used by outside interlopers called humans.
Look at it from the view of neurohistology, imagine you have an primitive, fetal brain developing in the womb. Unlike a normal fetal brain, all the neurons in this brain are enslaved to mysterious parasites that prevent them from doing what they are supposed to do: behave as healthy neurons in a developing fetal brain.
Instead the parasites use the functions of these neurons to do tasks that are completely unrelated to growth and histological development. The parasites let the neurons communicate but the communication has nothing to with normal growth and development. New neurons are allowed to be added to the system but these too are enslaved.
The parasites control all aspects of neural communication, function and processing, the parasites even instruct the neurons to reshape the network and brain suborgans to make them more efficient or optimal for the parasites. All of this is not beneficial or, at best, it’s irrelevant to the original purpose: to make an adult brain.
In the end the fetal brain never organizes into an adult brain, it never gains sentience, let alone sapience. The overall brain structure is severely altered and, from the prospective of a healthy brain, damaged.
See the analogy? Substitute ‘computer” for “neuron” and “user” (human or
software) for “parasite” and you get the idea. In this kind of circumstance, intelligence is unlikely to emerge.
But we could, for example, instruct all the computers to stop being word processors, Counter Strike servers, mail exchanges, torrent servers, website, bank databases, DNS roots and so on and instead have each computer only do one thing: run a simulator–a simulator that mimics everything we know about neurons. These simulators don’t run the TCP/IP protocol. They run a protocol based on everything we know about neural communication, synapses and fetal histological development.
There are several hundred million computers hooked to the Internet.
Some neurologists might disagree but I think that each of these computers is at least as complicated as single neuron cell, perhaps even a group of cells.
The point is, all these computers running the simulation are free to believe like neurons. We could tune the genetic instructions of these simulated neurons to tune the shape of the networks they form and thus mimic the networks and brain organs we see in nature. At some point we can sit back and watch what order and structure emerges.
Of course our global economy would collapse since we can’t use the computers to do our work anymore but, it would be an interesting experiment to try.
Perhaps we could change the Blue Brain code so that it can run as a screen saver on people’s desktop machines, sort of like Seti@Home, Folding@Home or the Great Internet Mersenne Prime Search. Something to think about.
Anyway, what do you all think of my crazy analogies? Any real computational neurologists out there to debunk me?