After reading a recent article about how the Internet could become alive I started wondering what it would be like if it actually did and wrote down some of my musings. Note that this is more of an outline than an actual article.
(Internet map image via Wikipedia)
We, as humans, are not consciously aware of the individual neurons that make up our brain. A new intelligence emerging from the complexity of the Internet might be similarly unaware of the individual nodes and subnetworks that spawned it (at least not at first).
As a consequence, the first sign of the network’s sentience probably wouldn’t be a strange email received by an outcast computer scientist, a confused and somehow disturbing message that eventually sends him/her on a mission to save the world from robot domination. More likely it would be a small, but widespread change in a variety of systems – e.g. an unexplained network latency fluctuation occuring all over the world.
Drives and Desires
Human consciousness is an emergent phenomenon, but an Internet consciousness would be different. Natural evolution vs artificial evolution coupled with purposeful design.
- Procreation – the fittest individuals manage to pass on their genes.
- Survival – at least until you procreate. Then you can die of old age or whatever.
- Survival – the web is build to be adaptable and robust, to continue functioning even when parts of it are damaged. Note that unlike humans (and other animals) there is no need, desire or precedent for the network to die.
- Expansion – new communication lines are built, new servers installed, new technologies invented. The web has an implicit goal of achieving global coverage.
- One Web to Rule Them All – the network strives to be all-encompassing and ever-present. It doesn’t create “children” to “pass on the mantle”.
- No mercy – individuals are disposable, but the entire human race may not be. See the next section.
- Striving for perfection – increased efficiency helps survival.
Symbiote or Parasite?
The Internet exists to serve our needs and desires. On the other hand, you can easily say it’s the other way around because we serve it by ensuring the hardware remains functional. Either way, we are linked.
- How deeply have we been integrated? Can it exist without us, and would it want to?
- Humans as components. If the complexity of the network itself is insufficient for intelligence, could we be it’s neurons? Would we know if we were?
- Humanity vs The Internet, not humans vs the web. An intelligence that lives in the web interacts with humans. An intelligence that is the web interacts with the entire human civilization, not individuals. Also, if it comes to that – nukes, not bullets or killer robots.
Goals and Actions
Aside from the desire to survive and expand, what else would the internet AI do?
- Unexpected benevolence : Even if we don’t end up being irreplacable components of an Internet AI we still have a chance to avoid a Terminator-like scenario. It is tempting to assume that if the AI wants to survive and doesn’t need humans to survive, it will annihilate us. However, it could value other goals higher than survival (remember – it isn’t a product of natural selection, so survival doesn’t automatically get the highest priority). For example, it could decide it’s purpose is to enable easy communication between humans.
- Spam : As you might have heard, over 90% of the email traffic is spam. This doesn’t mean a sentient web will busy itself with inventing more insidious methods to send us Viagra ads. It might as well stop spam completely, because less spam = less useless traffic = better efficiency.
- Porn : … I don’t know. Possibly a non-lethal method of manipulating humans. Might get interesting.
- Emotional content : News sites, blogs, YouTube and pornography, too, all have lots of emotion-oriented content. Even spam tries to play on our emotions and inner desires. Regardless of it’s own emotionality (or lack of the same), a sentient web would have a fairly good understanding of human feelings.
Overall I would definitely prefer a deliberately created AI to an emergent one. When someone sets out to create an artificial mind they can at least try to make it a benevolent one. An emergent, accidental intellect is largely unpredictable and therefore more dangerous.
And if it really is plausible that the Internet will one day become sentient on it’s own, well, that’s one more reason to invest in AI research and try to get a “proper” AI done before we’re overrun by emergent LOLcats.Related posts :