The Risk Factor

As I was reading the New York Times book review section this morning, I came across a review of David Levy's book, Love and Sex with Robots: The Evolution of Human-Robot Relationships (Harper/HarperCollins Publishers, SBN: 9780061359750; ISBN10: 0061359750, 2007).

Quoting from the review:

"Humans, Levy writes, are hard-wired to impute emotions onto anything with which we’re in intimate contact, to feel love for objects both animate and inanimate. And robots, he argues, might turn out to be even more lovable than some humans. By 2025 'at the latest,' he predicts, 'artificial-emotion technologies' will allow robots to be more emotionally available than the typical American human male. 'The idea that a robot could like you might at first seem a little creepy, but if that robot’s behavior is completely consistent with it liking you, then why should you doubt it?'

The review, by Robin Marantz Henig, a contributing writer for The Times Magazine goes on in its concluding paragraph:

"Levy spends so much time laying out his logical arguments about how and why we will fall in love with robots that he gives short shrift to the bigger questions of whether we would really want to. I’d have liked a little less gee-whiz, and a little more examination about whether a sexbot in every home, a Kama Sutra on legs that never tires, never says no, and never has needs of its own is what we really want."

This book should provoke some interesting discussion. Robots that have are more emotionally available than the typical American human male by 2025? How about French or Italian men? Is that 2030? I guess I'll have to get the book to see what Levy says about the emotional availability of the typical American woman.

Maybe the idea of creating future sex robots can help get students interested in taking up computer science at Cambridge University again.

As first reported yesterday in the Register and then picked up today by ComputerWorld, Microsoft has had to pull the plug on its on-line "artificial-intelligence Santa bot" that was meant to talk to children about what they wanted for Christmas. Seems that the bot, as ComputerWorld put it, "wandered off topic" when certain words - like pizza - were used.

According to ComputerWorld, "Microsoft recently added the artificial Santa as a bot that Windows Live Messenger users could insert into their IM buddy list as northpole@live.com."

You can read about the bot in a Microsoft press release I found from last year titled: For a Jolly Good Time, Chat With Santa on Windows Live Messenger. A line in it is: "Filling Santa in on Christmas wishes and asking all about how the reindeer are doing or what’s new at the North Pole are a few of the things kids can talk to Santa about. Santa can even tell kids where they stand on his list: naughty or nice."

I guess the press release forgot to mention that Santa would be informing the kids about whether he was naughty or nice this year.

Microsoft said in a statement posted on the Register site: "Yesterday we received reports that the automated Santa Claus agent in Windows Live Messenger used inappropriate language. As soon as we were alerted, we took steps to mitigate the issue, including the removal of language from the agent’s automated script."

"We were not completely satisfied with the result of these actions, and have decided to discontinue the automated Santa Claus agent. We apologise for any offence or upset caused by this disturbing incident."

I guess Microsoft tested this year's Santa bot using the same strategy it does on most of its products - let the users find the bugs.

The New York Times this week had an article on smart cars and how one will "soon" be in a showroom near you. It quotes Dr. Sebastian Thrun, a computer scientist who heads up Stanford's Artificial Intelligence Lab, as saying, "Within five years, it’s totally feasible to build an autonomous car that will work reliably in several limited domains."

Furthermore, the article says, "In 20 years, Dr. Thrun figures half of new cars sold will offer drivers the option of turning over these chores to a computer, but he acknowledges that’s just an educated guess. While he doesn’t doubt cars will be able to drive themselves, he’s not sure how many humans will let them."

It will be interesting to see what happens when the first smart car crashes into one driven by a plain old human driver and results in a severe injury or death. Will the smart car's software be blamed? Will the argument be that the human driver has to be at fault since the smart car is assumed to be more carefully driven? And will the case be argued by "smart lawyers," a term that seems somehow oxymoronic to me?

In historian Felipe Fernández-Armesto's survey book Ideas that Changed the World, there is a section entitled "Impossibilism." In it, he reviews some of the paradoxes that philosophers like William of Ockham’s raised for contentious debate in the 14th century, such as “God can order you to commit murder” or “God can reward good with evil.”

If William of Ockham were alive today, he would probably coin something appropriate about Microsoft’s problem reporting.

As I noted a few weeks ago, Microsoft captures and analyzes those errors that unfortunately but not unexpectedly pop up every so often, which on some days provides Microsoft with 50 gigabytes worth of problem data.

I was recently sent a link to a screen shot of an error message that I have never encountered:

Windows Problem Reporting Has Stopped Working
A problem caused the program to stop working correctly. Windows will close the program and notify you if a solution is available.

As the comments at the link note, this error message poses some very interesting philosophical paradoxes and implications. For instance, how can a solution be sent if the problem reporting scheme is not working? How can a solution even be available if the problem is not reported? Or does it really indicate that Windows has developed HAL-like self-awareness? This could help explain Microsoft's Potty Mouth Santa.

All this made me wonder whether:

a) Microsoft has another error monitoring program to watch for when its Window’s Problem Reporting code has an error, and whether there is another one to watch for that one to have an error, and so on: all this watch watching might explain why its operating systems are so large, and;

b) if (a) above is not true, does the Microsoft error analyst team have a category for this specific types of error, waiting in hopes of an error turning up some day indicating that in fact Windows is now self-aware, kind of like the SETI folks do in waiting for that special signal from space to appear?

According to the London Telegraph, "flirting robots" are invading Russian dating websites with the aim of gaining personal information from unsuspecting victims. CyberLover is one such robot that masquerades as a person seeking love on-line, according to the story. It interacts with a potential victim asking questions like, "When's your birthday? Where can I send you a Valentine's Day card?" and so on. The fear is that these robo-lovers could soon be invading popular social networks phishing for information.

I wonder what happens when one robo-lover encounters another on-line? Do they exchange code words so they know that the other is one of their own? Or do they just keep chatting one another up forever?

In today's Wall Street Journal, there was a note regarding a story that is in Seed science magazine regarding the question:

"If aliens are out there, how should Earthlings go about getting in touch with them?"

"The question has provoked arcane but furious debate among scientists searching for extraterrestrials. Because scientists haven't picked up signs of alien life near Earth, the debate is essentially philosophical, revolving around such issues as who rightly speaks for humanity and whether humans want to draw the attention of possibly hostile life forms."

"A dispute erupted recently among scientists over an effort to draft a protocol for messages going from Earth into space, reports David Grinspoon in Seed, a science magazine. Several scientists who believe that governments and other scientists should be consulted prior to any space-bound communications resigned in protest from a prominent study group on extraterrestrial intelligence."

This got me to thinking about my earlier post on Microsoft's error reporting, and my joking reference about it possibility being a search for artificial intelligence. However, what happens if a computer does indeed become self aware? Who speaks for the human race, and does the first self-aware computer speak for ones that come after it?