The Autonomous Autobaiter


Experienced baiters please skip to the section below titled "Introduction to Autonomous Autobaiting".

Introduction to Scammers and Scam Baiting.
There are tens of thousands of people in foreign countries sending scam emails where they claim that they have many millions of dollars tied up in a trunk at some security company. There are lots of variations to the scam, but one example is that this money was left to them by a relative, and being in a refugee camp there is no way that they can get to it.  They need your help in posing as a relative, or business partner or whatever.  You are to put their $30 million in your account, and keep 25% of it for your effort and help them get to the US to live a good life with their millions.  Great.  Let's do it.  But after starting an email correspondence with them, you find that you have to be given power of attorney.  That costs only a few thousand but it's worth it.  But then you have to pay storage costs - a few more thousand. Then you have to pay for a government document. It goes on and on until you have exhausted all your retirement plan in order to gain several million dollars that never existed in the first place.  Last year these scams have bilked people all over the world out of at least half a billion dollars.  Most people are too embarrassed to report it. The US and foreign governments won't do much about it because, individually, the amount is often too small.

So we have to take matters into our own hands.  There are a number of people counter-scamming the scammers, and leading them on with a feigned interest, and generally wasting their time as a pretend victim.  They are called baiters.  For example a baiter might send the scammer to the Western Union several times to pick up a money transfer that was faked. There are a number of ways of slowing down scammers using baiting, but one very efficient scheme uses a computer program to aid in generating email replies.  This is called autobaiting.

Introduction to Autonomous Autobaiting
Many effective autobaiters need almost no interaction with a user who is the baiter.  The user never needs to read the incoming email or tune up the outgoing response. These might be called autonomous autobaiters. Other very effective autobaiters require some user interaction, and this results in more effective baits that keeps each scammer tied up longer.

The goal
My goal is to develop a simple to use autonomous autobaiter with a good deal of artificial intelligence so that long effective baits can ensue, but still require a minimal amount of interaction with the user. The autobaiter is controlled by a Graphic User Interface (GUI) which forms an AutoBaiting Interactive Environment (ABIE). The GUI consists of a number of parts such as tables that capture the personality of the bait and provide diagnostics that aid in debugging the "social engineering". 

In the current version, the intelligence of ABIE is a bit weak in places.  These weaknesses can be best understood and repaired through extensive field trials. So rather than trying to solve all problems before launching a field run, an intermediate step with some user interaction has been adopted. This strays from the totally automous aspect of the ultimate goal. 

User interaction is in the form of "operator alerts", where a simple yes/no dialog box pops up and informs the user that a simple decision must be made. The alert is phrased so that the most probable correct answer is a default yes and can be quickly dismissed. Other operator alerts have pull-down menus with choices.
Operator alerts occur fairly infrequently and result when sparse or ambiguous information is given by the lad, or there is a crucial misspelling or poor punctuation. As experience with the system is gained, and upgrades are made, the operator alerts can be, disengaged or hidden and always be automatically provided with a "yes" decision. This version would achieve the ideal in autonomy.

(Footnote: The term "operator alert" comes from robotic manufacturing. When the artifical intelligence of an assembly line senses that it has failed, it calls an "operator alert" (often a red blinking light) for a human decision.)

Gleaning information
The most difficult problem in autonomous autobaiting is gleaning vital information from the scam email, for example the lads name. 
An operator alert is given if there is too much uncertainty.  The name becomes "Friend" if the user ignores the alert. The user can override that decision with a pull down menu of possibilities for which the autobaiter could not gain confidence.

Gleaning the proper response email is also difficult.
The correct email address for the response is often not the one used in the scammers first mailing. The lads often will suddenly change their email address because they were shut down for their criminal activities. Detecting that the new address comes from one of the lads already in the hundreds of history files is tricky.

A Nightmare
Telephone numbers, account numbers, the plethora of fake lottery numbers, etc. must also be handled. These concepts result in a nightmare of programming. The logic is often heterogeneous, and fuzzy.  It must be flexible enough to garner information, often poorly presented, in a wide variety of forms, but not so flexible as to interpret the information incorrectly. The methods which determine the lads name and alternate email are long and

Important Reason for an Intelligent Autobaiter
Successful autonomous autobaiting has been done with simple programs or scripting languages. Often these programs and methods can handle a massive number of scammers. In the overall goal, this is very effective in wasting a lot of man-hours of scammers time, so why stray from the simple and effective methods?  The major reason for adopting the difficult and lengthy intelligent design philosophy is that the program coding is being done by a retired person who spent over 25 years in a certian field of AI and now has way too much free time on his hands.
One personal challenge is to see just how far the computer can fool the lad. The program evolved over the course of six months -- a task that was often fun and more often very frustrating.  Hopefully a useful program and tractable user interface has resulted.

A second development goal was to use platform independent software that was fast and versitile.  The Java language was used for that reason. This program will eventually be thrown to the winds of open source.  A complete GUI
integrates and handles the autobaiting system which includes tables of keywords and scripts, a scam type classifier, history files of scammer data, a journal of the email exchanges over time, transmitting and receiving POP and web email accounts, and debugging modes.  An eventual goal is to separate the tightly coupled integrated system and attempt to form more stand-alone class definitions with simple interfaces that enable an SDK (software development kit) so that the system can easily be picked apart by other programmers.