Soon, Mobile Phones Will Spend Money On Your Behalf
from the disaster-waiting-to-happen dept
When the internet first started to really take off a decade or so ago there were plenty of stories about people working on so-called “autonomous agents” that would sit on the network and learn from what you do – to the point where they would be able to do stuff for you. The idea was they would create true “virtual assistants” that would, for example, see that you wanted to fly to New York City, and would automatically go out, buy you the best plane ticket and arrange for a hotel as well – probably while communicating with the opposite agent from the airlines and hotels. This vision died a pretty quick death for two big reasons: (1) the technology never worked very well (the best that came out of it was used for collaborative filtering networks to recommend what books you might like to read) and (2) people did not want these little agents doing things for them. They were scared that the agent would spend their money in bad ways, or make bad decisions. When it comes to your own money, most people want to have a pretty all-encompassing say in how it’s spent. So, I’m somewhat surprised to hear that a new group of researchers is trying to resurrect this sort of research with the ever popular twenty-first century twist of putting it on the “wireless network”. Some researchers think they’ve come up with the brilliant idea to create virtual agents that sit on your mobile phone, learn what you do, and learn to anticipate what you want, to the point where they may go out and buy stuff for you. This sounds exactly like the autonomous online agents from a decade ago, and I doubt very much has changed with regards to human nature when it comes to letting a computer spend your money.
Comments on “Soon, Mobile Phones Will Spend Money On Your Behalf”
Autonomous Agents are not dead yet!!!
Most of the work on autonomous agents has been done in the field of security, in autonomous network defense circles, and there are no plans of them dying anytime soon. Unlike your definition of an autonomous agent, the security autonomous agents don’t gather information, but instead look for strange anomolies (attempted hacks, etc.) on the network and modify the network accordingly.
While these agents are still a long way off, most of the work done outside of classified circles (obviously we don’t know where the government has gone with this stuff,) has been fruitful, though the technology to gather and review the data collected/actions performed has been slow to come to fruition. The problem really boils down to everything it has boiled down to in the past, information overload kills every time…
Re: Autonomous Agents are not dead yet!!!
Good point. I wasn’t suggesting that all “autonomous agents” were dead… just the kidn that would act as your “virtual assistant”.