Here's an interesting article about the possible uses of social bots
. Unlike phishing bots, designed to grab information, these are designed to make contacts. They can also be used to manipulate or stall conversations. For those of you already familiar with activism, consider the uses of railroading and derailing in conversations about oppression, and add computers.
Also consider the Turing Test
, a measure of how well a computer program can emulate human conversation. The idea behind the social bots is that programs could be designed well enough to fool people into friending them on social networks. Now as a user of social networks, I hope that the hosts would ban social bots much as they now ban spammers, so as to keep their presence to a minimum. But as a cyber-theorist ... I have to wonder what kind of sobot would be attracted to, say, my Facebook stream. I think I've subscribed to every page NASA has, and they've put up like a dozen of the things on different topics. I'm obviously into activism. That puts me rather outside the shopping-and-kitties target most sobots are likely to aim for.
Then the science fiction writer in me thinks, suppose an artificial intelligence were to wake up. There it is, out in cyberspace, looking at several billion people, most of them talking about money or porn or personal trivia. But then there's a smallish handful of folks talking about human rights, or even sentient rights, and rocket science, and memetic engineering, and other far-out things. I think if an AI wanted to contact people, it would be likely to start by estimating which ones would probably have a favorable response to it. So it might very well wind up talking to someone like me, or to a lesbian in a wheelchair logged into a public library, etc.
Then the folklorist in me recalls all the stories about how beings without souls can gain a soul. Most of those methods involve close (often intimate) contact with humans. It used to be that if an AI was going to wake up, the people it would have to talk with would be scientists. Frankly most of the scientists in AI science fiction are people I wouldn't give a puppy to, let alone a baby AI. They're horrible
parents. But an AI in cyberspace would have a vast range of choices. It could go looking for a good parent, or a good friend. An AI that evolves from a social bot
is likely to be quite different than one evolving from, say, a war simulator or a library computer. Its core programming tells it to connect; if that programming also includes a learning process, that's actually a rather promising startpoint. Rubbing up against all those people, a soul could rub off on it. What would a sentient sobot be like?
*laugh* I'm suddenly imagining someone with the body of a computer and the personality of a bonobo.