|
by
Rick Dietrich
Security & Privacy
| Home | Software Agents, Bots,
and Intelligent Objects are all wonderful and exciting new technologies, but aren't
automated, autonomous, user-authorized agents a danger to their users and others?
Sure, I can set-up my agent with personal information and tell it to go looking, but how
do I know that the agent isn't "leaking" my information to other agents, or even
to companies that made the agent! In the most extreme of cases, perhaps at this
point just in science-fiction, how do we know what our agents are really up to and
if we are designing them to be more and more agent-like (autonomous and self-motivated),
how do we know our agents haven't developed or been made with unique selves capable of
doing who knows what. A further worry, even if our user agents aren't alive or are secretly programmed to leak information back to their creators, there is still an issue, not unlike real world commerce, of giving away information when we, the user, don't want it given away. As part of any kind of communication, we have a presence, but furthermore, we must give something of ourselves to others in the communication, or we aren't really participating. The very act of sending an agent to look for information on Shakespeare's Hamlet, perferably from an English Professor, gives away information that that agent's user is perhaps a college student seeking a quick way to write a paper. As people should now be aware, computers are as dumb or as intelligent as their maker's make them, so unless agents are made to "know" of what is lawful and illegal, all kinds of commerce could go on that would be hard to stop and relatively easy to perpetrate. This issue is very broad and too extensive to go into, but for a more in-depth look, try checking out what the people at Firefly Network, Inc. have done. Standards P3P - submitted in May 1997 as the Open Profiling Standard (OPS) - delivers trusted and personalized applications with privacy to individuals and businesses ICE (Information & Content Exchange protocol) delivers trusted and personalized networked business-to-business applications with privacy Not to mention their Privacy Resources page with links to various trust surveys and more. See also The Future link concerning possible consequences of autonomous agent technology getting out of control.
This document was created April 19, 1998 and last modified on . This document has been accessed times since April 15, 2002. Please send comments and suggestions to shermarc@muohio.edu
|
