Trade.Information

Topic Name: Researchers say Software can now analyze your e-mails

Category: Computer science & technology

Research persons: Gilbert Peterson

Location: Air Force Institute of Technology, Ohio, United States

Details

very soon big brother will be able to follow you to work. Software is being designed to allow companies to flag up employees who are potential saboteurs, industrial spies or data thieves. It might also flag up whistle-blowers.

US companies surveyed earlier this year said at least one-third of damage to business due to cybercrime was committed by insiders. “Many of the biggest financial losses tend to be due to trusted insiders, individuals who steal or who disable computer systems,” says Gilbert Peterson at the Air Force Institute of Technology (AFIT) in Ohio.

Writing in a forthcoming edition of Digital Investigation, Peterson and colleagues say their software is based on an open-source algorithm called Author-Topic. Developed by researchers at the University of California, Irvine, it gauges which topics authors commonly write about. Fed a series of documents, such as academic journal articles, Author-Topic examines the frequency with which words appear in each and uses that to infer which topic that document is about. It then identifies topics that each person writes on most.

Peterson’s team uses the software to analyse emails, rather than articles, and extra software records whether people are sending emails internally or externally. Their system identifies people who are not discussing certain, expected topics - say social activities - with their colleagues, and flags them as possibly feeling alienated. It also identifies those who are discussing sensitive topics externally and classes them as having “clandestine, sensitive interests”. People who are flagged in both categories could pose a risk to a company, say the authors.

In addition to potential saboteurs, the software can also spot whistle-blowers. When it was fed the 250,000 emails sent between employees at bankrupted energy company Enron, it flagged employee Sherron Watkins as one of just three who were both alienated and had clandestine, sensitive interests. It was Watkins who blew the lid on the firm.

The search engine IDOL, made by Autonomy in the UK, can also detect insider threats, according to managing director Mike Lynch. But the AFIT system will be open-source, so organisations will be able to use it for free. In most US states such software is legal, but Ian Brown of the Oxford Internet Institute says that in Europe employees can only be monitored if they are suspected of fraud.

Note for Cybercrime

Although the term cybercrime is usually restricted to describing criminal activity in which the computer or network is an essential part of the crime, this term is also used to include traditional crimes in which computers or networks are used to enable the illicit activity.

Examples of cybercrime in which the computer or network is a tool of the criminal activity include spamming and criminal copyright crimes, particularly those facilitated through peer-to-peer networks. 
Examples of cybercrime in which the computer or network is a target of criminal activity include unauthorized access (i.e, defeating access controls), malicious code, and denial-of-service attacks. 
Examples of cybercrime in which the computer or network is a place of criminal activity include theft of service (in particular, telecom fraud) and certain financial frauds. 
Finally, examples of traditional crimes facilitated through the use of computers or networks include Nigerian 419 or other gullibility or social engineering frauds (e.g., hacking "phishing", identity theft, child pornography, online gambling, securities fraud, etc. Cyberstalking is an example of a traditional crime -- harassment -- that has taken a new form when facilitated through computer networks. 

About Researcher

Peterson, Gilbert L. 
Assistant Professor of Computer Science 
Dept of Electrical and Computer Engineering
Email: Gilbert Peterson 
Education:
2001 Doctor of Philosophy in Comptuer Science, University of Texas Arlington
1998 Master of Science in Computer Science, University of Texas Arlington
1995 Bachelor of Science in Architecture, University of Texas Arlington 
Research Interests:
Dr. Peterson's research interests include artificial intelligence, autonomous robots, multi-robot systems, digital forensics, steganalysis, and machine learning.