Given the level of the debate on climate change in the comments of blogs and on twitter, I am not surprised that robots can easily pass the Turing test in there. A 2010 report, reproduced below, tells of a pro-science chatbot, but I am sure that there are anti-science chatbots as well. Isn't it nice? So much hassle avoided! In the end, I hope they'll come back telling us who won. Wouldn't it be a good idea also to have drones fight wars among themselves for us?
Chatbot Wears Down Proponents of Anti-Science Nonsense
By C. Mims - from the MIT technology review
When he tired of arguing with climate change skeptics, one programmer wrote a chatbot to do it for him.
Nigel Leck, a software developer by day, was tired of arguing with anti-science crackpots on Twitter. So, like any good programmer, he wrote a script to do it for him.
The result is the Twitter chatbot @AI_AGW. Its operation is fairly simple: Every five minutes, it searches twitter for several hundred set phrases that tend to correspond to any of the usual tired arguments about how global warming isn’t happening or humans aren’t responsible for it.
It then spits back at the twitterer who made that argument a canned response culled from a database of hundreds. The responses are matched to the argument in question – tweets about how Neptune is warming just like the earth, for example, are met with the appropriate links to scientific sources explaining why that hardly constitutes evidence that the source of global warming on earth is a warming sun.
The database began as a simple collection of responses written by Leck himself, but these days quite a few of the rejoinders are culled from a university source whom Leck says he isn’t at liberty to divulge.
Like other chatbots, lots of people on the receiving end of its tweets have no idea they’re not conversing with a real human being. Some of them have arguments with the chatbot spanning dozens of tweets and many days, says Leck. That’s in part because AI_AGW is smart enough to run through a list of different canned responses when an interlocutor continues to throw the same arguments at it. Leck has even programmed it to debate such esoteric topics as religion - which is where the debates humans have with the bot often wind up.
“If [the chatbot] actually argues them into a corner, it tends to be two crowds out there,” says Leck. “There’s the guns and God crowd, and their parting shot will be ‘God created it that way’ or something like that. I don’t know how you answer that.”
The second crowd, Leck says, are skeptics so unyielding they won’t be swayed by any amount of argumentation.
Occasionally, the chatbot turns up a false positive - for example, it has a complete inability to detect sarcasm. This proved to be a problem when a record heat wave hit L.A. last summer, causing innumerable tweets of the form “It’s 113 degrees outside - good thing global warming’s a myth!”
Leck always apologizes when AI_AGW answers someone who isn’t actually arguing about the science of climate change and then subsequently whitelists his or her account. The bot also has a kind of learning algorithm in it in that can be trained not to respond to phrases that cause false positives.
In the future, Leck would like to expand AI_AGW by giving it the ability to learn new arguments from the twitter feeds of others who debate climate skeptics - allowing it to argue into the ground an ever expanding array of anti-science tweeters who are unwilling or unable to look up the proper scientific literature themselves.
In a way, what Leck has created is a pro-active search engine: it answers twitter users who aren’t even aware of their own ignorance.
Update: some guy on Hacker News sums it up better than I ever could.
No comments:
Post a Comment