Automated Untelligence Application

Automated Untelligence - Dictionary Edition

[atsimg]http://files.abovetopsecret.com/images/member/149d649182e6.jpg[/atsimg]
After being inspired ATS member IgnoranceIsntBlisss in this thread, I decided to spend a few hours today putting together a simple Automated Untelligence Application, and I figured it deserved a thread of its own. This one doesn't scrape the internet for words and phrases however, instead it uses it's own dictionary file stored on your computer. You place one word per line in a text file and search queries will be formed from words chosen at random from your text file. You could probably also have one sentence fragment per line to create more coherent search queries, I haven't tried it. I've been using a massive text file with essentially every word in the English language and more, but I've found a problem with that.

A large portion of English words are large and not widely used, therefore a lot of the time the search results will bring up other dictionary files and things like that, and if your searches keep bringing up results like that, the search engine will probably recognize that you're just using random words from a dictionary. The answer would be to use a list of more common words and sentence fragments, I could have built an engine to scrape data from the internet and create a text file like that for you, but I just don't have time. The help section hasn't been done either. If anyone wants the source code to add such a function or any other functions, PM me. It's written in VB2008 so it's very easy to build upon.

[atsimg]http://files.abovetopsecret.com/images/member/4781e4515b1b.jpg[/atsimg]

Functions:

- Minimizes to system tray so it will run in the background unnoticed.
- Includes the "CHA0S Reactor" which will generate random text fragments from your dictionary file.
- Search URL in the settings can be modified, meaning it works with most search engines, not just Google.
- Time between searches can be modified (meaning less CPU usage and less chance of the search engine banning you).
- Words per search query can be modified.
- External dictionary file can be modified from within the app.
- List of valid symbols to use in search queries and text generation can be modified.
- It can also act as a general internet search tool.
- No installation required.

Requirements:

- Windows XP, Vista or 7
- .NET Framework 2.0 or higher

Download Application

Download Dictionary - This is the dictionary file I was using, as I said, it's not very useful.

Warning: I have not used this application for long periods of time, I will not be held liable for any damages this application may cause.

This picture doesn't really prove much for a skeptic, but I give my assurance that this application is totally virus free, hopefully others can scan it and verify that.

[atsimg]http://files.abovetopsecret.com/images/member/f83059d25ac5.jpg[/atsimg]

EDIT: Bug fix applied. The spaces in sentence fragments will be converted to + symbols if you're using sentence fragments instead of one word per line. Editing such a dictionary file from within the app will look clustered and hard to read. Open it in notepad and make sure you have one sentence fragment per line, the app separates items by the return/enter key.

[edit on 26/7/10 by CHA0S]

I kind of see what you are trying to do even without downloading but I think a batch script would probably be reasonably easy to do the same thing in fact I've just done it out of curiosity.

I'm sure you're program works fine though and no doubt people will find it more convenient but I think there is probably little point in it they can probably tell which sites you have clicked on anyway.

For fun here is the Batch Script I came up with I'm sure someone will be able to improve it easily enough it's been a long time since I've tried and I had to keep correcting little things until it worked right.

@echo off
for /f "tokens=* delims= " %%a in (dictionary.txt) do (
set /a N+=1
start iexplore google.com/search?as_q= + %%a
PING 1.1.1.1 -n 10 -w 1000 >NUL
taskkill /f /im iexplore.exe
)

To explain it a little, basicly it just reads the lines from dictionary.txt and opens a google search in intenet explorer then pings to a non existant address to delay it for 10 seconds it then closes Internet explorer and does the same for the next word in dictionary.txt.

Not trying to take anything away from your program it really just made me curious if I could achieve the same result and I'm also just relying on a dictionary file for search words.

[edit on 26-7-2010 by Teknikal]

reply to post by Teknikal


Cool, I have no idea how to write batch scripts, but that's interesting none the less. Does it take random words from the file and form a query or string or just search one word in the dictionary.txt file at a time? Glancing over it, it just seems to read through the file and search one word at a time.

EDIT: To clarify, my app takes random words from the text file, and forms a search string like: apple+bounce+whale+soup

[edit on 26/7/10 by CHA0S]

Yes unfortunately I only went for the next line approach though I'm sure someone else could do a lot more. The fact your program is combining words does seem to be a better Idea though so definitely a big advantage there I didn't even think of that.

Same thing could probably be done in a batch script but they do have a limit on usefulness and of course they can get complicated pretty easily so probably not advised. All in all I'd say your program has a lot of advantages though as I'm doing things like just killing Internet Explorer as a short cut.

Might have another go next time I have some time to kill.

[edit on 26-7-2010 by Teknikal]

Damn man you do all this and don't even tell me??

I haven't even looked too closely at what you've done yet...

reply to post by IgnoranceIsntBlisss


Haha, I made a post on your thread before I made this thread. I figured you'd notice it and there was no point PM'ing you.

BRAVO!

That's like exactly what I envisioned years ago!

You have the browser built-in and all!!

For that area it does exactly what it needs to do. It's actually more simple looking than I assumed it would have to be. You're an artist sir!

SeH pointed me to this page which might be helpful to people:
stackoverflow.com...
and
www.webmasterworld.com...

He brought up a concern that to rapid of searching might have Google block ones IP address, so a frequency setting might be ideal? Have you been running it for an extended period of time?

EDIT:

We're sorry...
... but your computer or network may be sending automated queries. To protect our users, we can't process your request right now.

See Google Help for more information.

Definitely needs a throttle! DAMN!

Google's maximum search string is 13 words last time I checked. if it varied how many words went in each time it could make it harder for google to detect the automation.

Are you going to release the source?

So many thanks!!!!


[edit on 27-7-2010 by IgnoranceIsntBlisss]

reply to post by CHA0S


Im sure the major search engines already have measures in place to prevent attacks such as this.

Unless there are a large number of copies of AU running, each copy is using the same dictionary and random number seed the number of searches for each distinct string is going to be too low to affect the AI.

But in this case the SE can predict the next string. So if the number of previous searches is low or the number of suspect sesrches is high and each word is in the known databases then discard and flag as suspect. Then log the ip and remove from search storage for 48 hours.

I know you only knocked this up but it would take a lot more effort to make a difference. And it would take a sustained war similar to the one spammers undertake.

Anyway, what is the reason for trying to thwart the AI?

[edit on 27/7/2010 by LightFantastic]

reply to post by LightFantastic

quiches ballrooms demurrer reroller raviolis

fourgons brokers preadmit pea wronger

metaphor waylaying redocked- chives fingerprints

sacrilegious? menially bounciest artiest apologizes

unthrone snottily plasmas caret dwellings

burghal deride fezzed- girdler venin

lashins stand surroyal annulose locoing

loneness decider faitours beyond serialisation

Those are the first 8 search queries I just had it do. I'd say his v1.0 dictionary is damned good. I couldn't have come up with anything near that selection.

I'm confident we're off to the right start.

Check out the 1st link in my sig for what this is about, but in short we're trying to slow down its march to AGI. Thanks tho, we need to figure out all of the flaws and strengths in these sorts of ideas. Also check out the other AU link in my sig. This tool is only one part


Basically it can only run at 1 search every 20 seconds, or something along those lines, instead of every second. And then eventually millions of people doing it, should put an impact in everything else everyone else is doing to help the system learn.

[edit on 27-7-2010 by IgnoranceIsntBlisss]

So your trying to delay the development of AGI by spamming googles search function?

I think you might need a bigger stone David!

reply to post by VonDoomen


The idea isn't just to hassle it and waste it's time, its to juggle word combinations in a non-sensical way to impair its ability to associate words and thereby gain understand in meaning and natural language processing. Check out the links, we're devising different methods to do it and not just in search queries.

reply to post by IgnoranceIsntBlisss


Oh i gotcha. Hmm i would say i still have to disprove of this action. i think AGI is a good thing. i mean, if its true AGI, i dont see why it would be against "us".

reply to post by VonDoomen


The technology itself isn't the immediate concern...

It's the despot technocrat immortalist plutocrat eugenicist social-darwinist control freak oligarchs that at the commanding heights (and are already robbing us blind):
www.abovetopsecret.com...
agimanhattanproject.com...

reply to post by IgnoranceIsntBlisss


You already know what I'm about to say since I sent you a PM, but for everyone else, I will answer these questions.

Definitely needs a throttle! DAMN!

Changing the "Wait each Search" setting should fix that. It's in milliseconds, so 1000 = 1 second delay each search.

if it varied how many words went in each time it could make it harder for google to detect the automation.

I'm working on v2.0 of AU - DE at the moment. It will include the ability to randomly vary the number of words per search to a degree adjustable in the settings.

Are you going to release the source?

Yeah, when I release v2.0 I'll probably upload the source. However, I will be working on a client project for the next week or so, and AU - DE obviously isn't my top priority. And I also intend on including an AI engine in version 2.0 which means it will learn and the sentences get better the longer it runs.

EDIT: Sorry, fixed up those messed up quotes.

EDIT: I am waiting on info from my client before I can continue any further with their project, so tonight I will probably be working on AU - DE and the only thing I really have left to do is the AI engine, I've done everything else I just talked about last night, and also lots of little bug fixes. If we're lucky I should be done tonight. I'm trying to keep it still simple but a lot more effective based on what I've learnt so far.

[edit on 29/7/10 by CHA0S]

AU - DE v2.0

[atsimg]http://files.abovetopsecret.com/images/member/27436de4b737.png[/atsimg]

Here we have it, AU - DE v2.0. It includes many new features and bug fixes. I also provide a much better dictionary file with this release. Please refer to the inbuilt help section before asking questions in this thread. I've also provided the source for anyone wanting to enhance or modify the application. Report any bugs on this thread.

Features:

- Version 2.0 includes an A.I. engine which will get "smarter" the longer it runs.
- A.I. engine includes numerous settings and can be disabled.
- Minimizes to system tray so it will run in the background unnoticed.
- Includes the "CHA0S Reactor" which will generate random text fragments from your dictionary file.
- Includes the Word Fusion utility - will join dictionary text files and remove duplicates.
- Search URL in the settings can be modified, meaning it works with most search engines, not just Google.
- Time between searches can be modified (meaning less CPU usage and less chance of the search engine banning you).
- Words per search query can be modified with a degree of randomization.
- External dictionary file can be modified from within the app.
- Can use custom proxy list or import a text file from the internet.
- Adjustable Time Out string. Helps detect when the search engine has banned you.
- List of valid symbols to use in search queries and text generation can be modified.
- It can also act as a general internet search tool.
- No installation required.

Requirements:

- Windows XP, Vista or 7
- .NET Framework 2.0 or higher

Downloads:

-Application
-Dictionary File
-Source

Note: I made this as quickly as possible, and so the code is not spaced or commented.

It is free to use and modify, but don't go replacing my name and avatar and pass it off as your original work.

[edit on 1/8/10 by CHA0S]

Breathtaking!

You've about wrapped up one of the major prongs of this overall ordeal. I hope that one day soon your effort here may serve as an inspiration to many.

My only concern with your new features is that the AI might actually feed it 'coherent' sentence structures, which would only keep Google busy instead of confused. Maybe I have the wrong idea about what that does?

Watching this thing go, it's so impressive!

I've provided an easier link for people to download the app+dictionary:
agimanhattanproject.com...

Ut oh:

"Error
Your text file is corrupt! You have most likely deleted or modified words in your dictionary file needed by the AI engine. The process will now stop. Word = digital"

[edit on 1-8-2010 by IgnoranceIsntBlisss]

reply to post by IgnoranceIsntBlisss

My only concern with your new features is that the AI might actually feed it 'coherent' sentence structures, which would only keep Google busy instead of confused. Maybe I have the wrong idea about what that does?

Well, I asked SeH the same question. Why is it good to have some form of structure in the sentences - because I thought the point of it was to be "untelligent" so it confuses their learning systems...but logically, that will only serve to strengthen their ability to distinguish generated queries from real queries, and by having a degree of structure in the queries it confuses them even more.

I explained it like so in the help section of the app:

This aim of this project is to slow down the growth rate of the ever increasing database of knowledge search engines like Google possess. Our ultimate goal is to interfere with their data-mining operations so vastly that it becomes much harder to produce useful and reliable data, and their output of such data becomes significantly lessened.

This is achieved by sending random nonsensical search queries to the search engine so as to confuse their learning algorithms. By generating slightly structured sentences, in this case, by using words found to posses a relationship, those search queries become harder to distinguish from real search queries, and thus are believed to be more damaging to their learning algorithms.

However, if left to run for long enough the queries could become quite structured, and that's why I've even provided a "throttle" feature on the A.I. Changing the setting "Use word relationships for queries 1 out of every how many times?" will achieve that, refer to the help section for further information.

[edit on 1/8/10 by CHA0S]

Originally posted by IgnoranceIsntBlisss
Ut oh:

"Error
Your text file is corrupt! You have most likely deleted or modified words in your dictionary file needed by the AI engine. The process will now stop. Word = digital"

Grrrr...I thought I had fixed that bug for good...it looks like this release will be pretty useless until I figure it out once and for all...there are a few spelling mistakes I missed as well...I'll update you soon...

PS - Did you save your dictionary file after that session? Could you send it to me if you did.

EDIT:

Bug fixed: Download the application again if you downloaded it before I made this edit.

EDIT: I'd thought I share this one line of code I used:

SrcFrags = Replace(SrcCop, Str & " ", "☻").Split("☻"c)

It has smiley faces!

[edit on 1/8/10 by CHA0S]

No i didn't save it.

So before the AI 'crashed', it ran for a good 12 minutes.

Now I ran it without the AI and Google stopped me after a minute or so.

I'm using:
om/ProxyList/fresh-proxy-list.shtml

... oh wait it just started going again while I had the proxy window open

went a few mintues and stopped me again.

[edit on 1-8-2010 by IgnoranceIsntBlisss]