Friday, January 6, 2012

How talking to your phone could change you

Both Android phones and iPhones have voice-controlled interface software, but their interfaces are very different. Android phones have Voice Search where the user speaks "commands" and the phone responds with results (or actions if programmed to do so). iPhones have the now-famous Siri where users "ask" the phone questions or requests and the phone responds in a human-like conversational style.

I'm a bit uncomfortable with using voice interfaces. I'm not sure why: perhaps it's because I don't want to sound like I'm talking to myself, or maybe I don't like the sound of my own voice. I'm really uncomfortable with the idea of talking to my phone or my computer as if it were a person. Once you start doing that I think it's a very short step before you start attributing real human characteristics to the machine.

Computer geeks are well-known for giving clever names to their computers and devices, but for me at least it's not a personal thing. Computers need names to help identify them on a network, and because computer geeks have multiple computers they like to use a naming scheme themed on something they have an interest in. I have named most of my past computers after monsters, and my current laptop is named 'wyvern'. However I have never actually talked to my computers the way many petrolheads are known to talk to their cars.

The episode of Radiolab I linked to below includes three stories of just how easy it is for people to start treating machines that pretend to be people as if they really were people, even when we know at a conscious level that they're machines. For me, the scariest story was the one in Clever Bots about Joseph Weizenbaum's secretary spending many hours after work "talking" to ELIZA about her personal life even though she was fully aware that ELIZA was not human. As someone who has difficulty with the art of human conversation, I can see how I could fall into a similar trap.

After listening to the program I felt that the science supported my concerns and I wasn't being overly paranoid. I support full discrimination between humans and machines. No rights for robots! And a big NO to human-robot marriage!!!

Source: Talking to Machines - Radiolab:

'via Blog this'

Will robo-wars become a cat-and-mouse game?

I think the current state of "robo-wars" is too imbalanced to be sustainable over the long term. The USA has launched over five times as many drone attacks against Pakistan (which it is not "at war" with) than manned bomber attacks in the Kosovo war 10 years ago, and the US military shows no sign of slowing down their use of semi-autonomous war machines.

It would be difficult for opponents of the US to match their level of spending and technology investment to build or acquire equivalent war-bots. However they might be able to find ways to turn these machines against their masters. This could vary from jamming a drone's control signals to take it down and convert them to the enemy's control, to planting malware inside a drone's control systems.

Drone malware doesn't need to be sophisticated to cause a lot of damage. One of the incidents mentioned in this radio program gives a clue on how this could happen. The malware could wait until the drone meets the conditions that suggest it is inside home base (e.g. a GPS location or receiving a control signal to dock) and then tell the drone to "go squirrely" (i.e. spin around in a circle firing all its weapons).

Autonomous and semi-autonomous war machines are amoral: they have no sense of 'loyalty'. It may be impossible to tell by looking at a robot (i.e. the hardware) that it's programming (i.e. the software) has changed. Soldiers are unlikely to be able to tell if and when one of their robots has been reprogrammed and is no longer 'trustworthy'.

As we have seen from Stuxnet, malware can be implanted at a very low level in embedded systems. When such malware activates, it can be extremely difficult to distinguish their effects from unintended software bugs.

Of course, once such war-bot malware appears, you can be sure that countries like the USA will work on developing countermeasures. A new technology cat-and-mouse game will then join the existing games of virus/antivirus and black-hat/white-hat.

Source: The morality of robo-wars: PW Singer - The Philosopher's Zone - ABC Radio National (Australian Broadcasting Corporation):

'via Blog this'

Indoors: Climate-Controlled Paradise or Toxic Gas Chamber?

Some of my work colleagues are a bit picky about the environmental conditions inside the office. The pickiest of them all is the group that I call "Management".

Management is very sensitive to bathroom odors, so makes a bit deal about using "air fresheners" to reduce the offensiveness of the smell. Management is also sensitive to temperature, and likes to keep the air conditioning set to a nice cool temperature. Opening the windows automatically disables the air conditioning, so Management usually keeps them closed.

Not much research has been done on indoor environments until relatively recently. One of the findings mentioned in this Future Tense program is that the chemicals in many air fresheners react to ozone to create toxic by-products. Most indoor ozone comes from outdoor air but is also generated by some equipment like laser printers. However once these toxic by-products are formed in a sealed indoors environment they tend to hang around for a while.

While these by-products are not being formed in high enough concentrations to cause office workers to suddenly drop dead all over the world, we don't yet know the health effects of long-term exposure to low concentrations of these by-products. For now, the best advise is to take it easy with the air fresheners and open the windows every once in a while.

Source: Indoor ecology - Future Tense - ABC Radio National (Australian Broadcasting Corporation):

'via Blog this'