Archived

This topic is now archived and is closed to further replies.

sakky

Security Systems with AI

Recommended Posts

I was discussing computer security systems and viruses with one of my friends earlier today. We came up with this crazy idea for a supreme security system. It would be one that defends it’s self by attacking its attackers to protect its self from hacks and viruses. We came up with a name for it—“Tyrant”. A voice activated and operated security system. It would identify a person by their voice as you would identify a person with their finger prints. Tyrant’s primary goal is to serve you like a secretary. It would load programs and play games once you spoke the command. To protect it’s self, it would use many hack tricks against hackers trying to break the system. When a hack is uncovered, it then quickly finds a solution to deal with the offenders efficiently. This would include hacking their systems or dropping in a deadly computer virus to destroy or shutdown their system, making Tyrant an efficient system killer. It would make use out of some of the best AI algorithms to aid in it’s destruction against prying eyes. Just like a Quake3 bot seeks and destroys so does Tyrant. I think it would be awesome to have an artificial intelligent security system at your side. You could talk to it and it would listen to you. Better then some other people I know. My friend and I then came to the conclusion, what if the damn thing turned on us? We would be screwed.

Share this post


Link to post
Share on other sites
Don''t you watch scifi? It would do something like decide that the internet traffic you are requesting is really an attack and take down that web server, and then the next you asked for, ad infinitum(sp?) until it ruled the world.

Share this post


Link to post
Share on other sites
Now, i know that most of you guys are going to think tha tI am crazy, and maybe I am, but me and one of my friends were talking about this at the time Blaster came out, and we came up with the same sort of idea, except our model had a "blocking" device that would not let it turn on us. Basicly, it couldn''t learn to much, it''s primary duty was to hack back at people and act as a proxy firewall. Our''s wasn''t voice activated though...

Anyway, I am a major computer security geek. I love the stuff... I check CERT at least once a day... like I said call me crazy... I also do think that some day in the near future something like this will happen. They''ve got to do something about proxy servers on a LAN though, the security for those is real weak. That would be the strong point of an AI system, it could strengthen its level of plociy denial whenever it saw fit. I think that an idea like this is very exciting... if you couldn''t already tell

Share this post


Link to post
Share on other sites
We can''t even teach a football to play football right... and football has rules! How are we going to have it stop hackers when there are no rules to hacking?

Dave Mark - President and Lead Designer
Intrinsic Algorithm -
"Reducing the world to mathematical equations!"

Share this post


Link to post
Share on other sites
There are already AI network monitors that attempt to identify illegitimate events on the network and discern whether these are security breaches or not. This sort of research has been going on for quite some time and there are several systems that I saw at PRICAI''00 that offered reasonable results.

None of these systems though attempted to attack the source. They merely cut it out of the network. With the advent of vLAN technology, this should be even easier to accomplish.

Timkin

Share this post


Link to post
Share on other sites
I don't think having the system attack the attackers is a great idea. What if someone were to get a program onto an innocent person's computer that was designed to attack that security system to direct any possible counterattack away from the one actually responsible? People could even go so far as to attack other people by putting such a program on their victim's computer just so that it would be attacked by the so called security system.

[edited by - SyntheticHate on October 9, 2003 5:03:55 PM]

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
Hi, having some problems (I''m at someone else''s computer and I forgot my password) so I''m posting as a guest. Anyway, if Tyrant was that smart it could lauch nukes remotely. That is BAD. On the other hand, there is always the OFF button... but Tyrant would copy itself to other machines and keep yours on with their power. They could launch attackls on web servers and commandeer the pages... We have Slammer Jr, guys. We need an infallible way to turn it off, because it would reprogram itself, mabye not to accept terminate program commands. We would either need an infallible OFF, or have it completely user controlled. On the other hand, what about evil users?


To make a long story short, make an infallible OFF.

Share this post


Link to post
Share on other sites
Well, to be serious: It would be illegal in many countries to make a program to atack other computers, or at least to use it. Also, what happens if somebody attacks you using a computer with the AI program on it? If it can learn, what happens is that you counter attack, they counter attack better, you counter attack better etc until both computers are entirely devoid of ''life''. I don''t know about you but I, personally, would not want some ''super AI'' on my pc fighting with another one somewhere.

Preventing access sounds like a MUCH better idea. It would also be good if there was a site for it to monitor to download info about new security holes so it could automatically turn off/block the ports of/etc problematic services/servers until a patch is released. Of course, that would only work if it ran on a gateway rather than desktop machines (don''t think people would like being told all the programs they use daily can''t be run because they are a security risk - need to do something else about that).

Share this post


Link to post
Share on other sites
So wait... what happens if Tyrant1 is protecting a college campus, and a student attacks corporation X that is protected with Tyrant2. What happens?? Do they duke it out? Maybe you could even make a web stream of virtual tyrant fighting. Sounds like a plan to me ^_^

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
I think we need a better name for a defense program than Tyrant. How about Guardian Angel, or GA? Why don''t we have GAs have virus defenitions, but also able to identify new viruses. When a GA finds a virus, it updates the virus defenitions of GAs all over the world. Also, if a new security hole is discovered, GA could reprogram out the hole, including if the hole was in GA itself. GA would notify other GAs if it was infected; GA would block all incoming data from known infected computers. If there was an epidemic all GAs would block all their incoming or outgoing internet transmissions.

Share this post


Link to post
Share on other sites