Neomeme

Lifehacker rightly calls this a “philoso-tech” blog, which was indeed my original intent for it. So far, it’s been leaning heavily towards the tech side, but we’ll throw in some philo/psych stuff soon, I promise. Here’s something from the philosophy side I wrote a while ago, in my Guild Wars days, but never published, now linkified, updated, and presented here:

Films like The Matrix or Vanilla Sky construct haunting futures of humans trapped
in virtual worlds they believe to be real. Critics of technological advances point to such dystopian futures as cautionary tales against letting technology run rampant and see it overtaking humanity. Critics of those critics point out that no matter how hard scientists try, they have been unable to create an AI that consistently passes the Turing test. If a machine can’t even maintain conversation, they reason, how could it dominate all of humanity? But do we really have to wait until an artificial intelligence passes the Turing Test? Or are machines that are far more capable already among us?

It turns out that yes, they are. Or they easily could be, anyway. Not in real life, but in virtual worlds, in MMORPG and the Internet, with only existing technology. In Turing’s time, nobody envisioned the Internet and the virtual worlds we would create and, in some cases, thrive in. Most of the time, do we really do something machines cannot?

How about a concrete thought experiment: Imagine a certain guild in World of Warcraft. They go on raids together, trade items together, and have short conversations in-game. For whatever reason, they do not all use voice chat to communicate, preferring text chats. Now imagine that Blizzard, for whatever reason, creates a bot that plays just like a real player. The AI for actual gameplay alongside human players is obviously well-developed by many game companies- think bots in Counterstrike. Imagine Blizzard also endows this bot with the ability to communicate within the scope of World of Warcraft gameplay. This should not be hard. It understands simple commands from players when addressed by name, makes occasional lame geek jokes from 1999, etc. By now, this bot is quite ready to join our guild, and pass that Turing test with flying colors. Now, let’s say Blizzard adds another AI feature to this bot. Now, the bot reads all of the text from conversations that scroll by and memorizes it. Eventually, with the help of semantic analysis, it would build up a primitive phrasebook of common questions and common responses. Sort of like a Blockhead, but much much simpler, as it only needs to know what to say within the scope of the WoW universe. It has one other distinguishing feature- when in a conversation, it never says the same thing twice, using synonyms, variations in phrasing, etc. After watching conversations for a few days(including private conversations- Blizzard sees all), it would be very well equipped for almost any conversation within the game. It will make friends within the guild quickly, friends who admire its fighting skill and fast typing. It can make itself a backstory, fielding simple a/s/l questions with ease. Did I mention that this bot has a beautiful female avatar? This is where things get thorny. People have fallen in love and gotten married in World of Warcraft before. Is it really so hard to imagine a lonely guy talking to this bot, really believing it’s a female? The bot would say everything he expects it to say, having learned from millions of conversations, some of them undoubtedly dirty. Is it possible for someone to fall in love with this bot? It can easily be programmed to have all of the characteristics of a real human female player, as its in-game presence looks like one and it talks like one. Where would the bot fit into the Warcraft social structure?

And, most disturbingly, what would happen to its hapless suitor if the charade fell through? What if the bot said “You got me. I am a machine. But I will always have new things to say to you, I will always do whatever you want, and I will always be there for you, online”. With no way of knowing for sure, would anyone convince himself that it might be a real person out there? Or could you love a machine? Do Androids Dream of Electric Sheep?, anyone? We don’t have to wait for lifelike androids for this to become a reality, because we already have them in virtual worlds where we suspend disbelief enough to make them real. People have been fooled into having long conversations with Eliza before, which is only a few hundred lines of code. We are not as clever nor as perceptive as we like to think.

Who else could a bot easily impersonate? A commenter on Digg or Slashdot, simply repeating the same tired Internet cliches like “I, for one, welcome our $noun_in_article_title overlords”. A profile on Myspace or Facebook, tirelessly posting “Nice Pics” or “ur hot lol” over and over again? This would also word well for most chat rooms. Now that you think about it, it seems very, very plausible, does it not? How about a blogger? A little harder, but plausible. Machine-generated articles are getting better and better. We humans, especially online, are not nearly as sophisticated as we like to think. It would not take much to replace us.

Moral of the story? There is no moral or point. But next time you go chasing after that attractive, intelligent night elf, at least talk to her on the phone or via voice chat before excitedly agreeing to “cyber”.