I was, of course, referring to AI-FW version 26.10 Which is endowed with
'artificial wisdom'. (after all the bugs are worked out) Traditionally &
annually scheduled to be next year. :) Seriosly, if we can use AI now for
things as serious & complex as medical diagnosis, why not security. It is,
after all, only there to "help". Not "decide". Think of it as diagnosing
your system. What changes it from rule based systems is the ability to
react to new & first time situations. To "adapt" to the unforseen. In
terms of the "social engineering" side of things, as you pointed out in an
earlier post, there may never be a complete solution. But, aside from
alerts & logs, it is difficult to police say 6000 users on 4000 computers.
For an AI agent, it would be a snap. Even autonomously changing policys &
permissions from time to time. It is easy to be afraid of this (which also
seems traditional) but sooner or later we have to admit that the
agglomeration of todays systems & networks is getting too complex for us
"mere mortals' to watch over 24/7 or even 16/6. Another issue with an AI
agent would be assisting in the "whole" security issue. Of which, firewalls
are just a part. Things such as vulnerability checks, automatic system
penetration testing (and reporting), intrusion detection & "reasoned"
response, timely educational warnings & updates to users interactively (as
they try to log on for the 6th time because they can't remember that new
password). Possibly even self checking it's own code for changes &
vulnerabilities. The technolgy to do all this is there now. This isn't
some new paradigm, but application of tried technolgy in a different area.
What do you think?
"Knowledge is a process of piling up facts; wisdom lies in their
simplification." -Martin H. Fischer
*-----Original Message-----
*From: [EMAIL PROTECTED]
*[mailto:[EMAIL PROTECTED]]On Behalf Of Gary Flynn
*Sent: Thursday, November 30, 2000 8:11 AM
*To: firewalls_list
*Subject: Re: Emily's response to the rumors
*
*
*Larry Paul wrote:
*>
*> The Universe produces them faster & faster. (us idiots) Security expert:
*> Yesterday I couldn't spell it, today I are one :>) Another reason for
*> considering AI in security.
*
*We can't code straight-line logic without bugs let alone "artificial
*intelligence" modeled after so-called "human intelligence" :)
*
*"If there's artificial intelligence, there's bound to be some
*artificial stupidity."
* --unknown (to me)
*
*--
*Gary Flynn
*Security Engineer - Technical Services
*James Madison University
*
*Please R.U.N.S.A.F.E.
*http://www.jmu.edu/computing/info-security/engineering/protecting_y
*ourself.shtml
*-
*[To unsubscribe, send mail to [EMAIL PROTECTED] with
*"unsubscribe firewalls" in the body of the message.]
*
-
[To unsubscribe, send mail to [EMAIL PROTECTED] with
"unsubscribe firewalls" in the body of the message.]
