Curtin University ‘spam busters’ are making tomorrow better by taking the fight to the illicit email and internet communications industry.
The Anti-Spam Research Lab headed by Vidy Potdar within Curtin’s Digital Ecosystems and Business Intelligence Institute researches the scale and cost of spam, and is devising ingenious ways to tackle it.
“We do research and try to develop new technologies that help stop spam and its associated problems such as fraud, scams, phishing attacks and cybercrimes,” Dr Potdar told Curtin News.
The lab started in 2008 and houses five researchers, as well as visiting professors, at its Bentley technology park headquarters.
One common question the lab is trying to answer is: ‘do spammers actually make money?’
“We’ve found that they do,” Dr Potdar said.
After achieving ethical approvals, the lab recently posed as a spammer on blogs and forums, creating a website to promote mock pharmaceutical products. Customer clicks were solicited via well-worn spamming tactics.
In a forthcoming paper authored by Dr Potdar, Pedram Hayati and Nazanin Firoozeh, the lab will reveal that 7769 spam posts attracted 2059 visits to the site – an impressive click through rate of 26.5 per cent.
Based on common costs per click paid by advertisers, the spammers stood to gain between $2059 (if $1 per click was assumed) and $411.80 (at 20 cents per click) – not a bad result from just 7769 posts.
More disturbingly, three of the unsuspecting 2059 clickers created a user account and tried to buy product from the site before the lab shut the transaction down.
In a worst-case, real-world scenario, the payments would have gone straight into the pocket of the spammer with no product being provided in return.
Pedram Hayati, a co-founder of the lab who is doing his PhD in anti-spam, explains.
“You need to make the anti-spam filters more and more complex to weed out spammers but then you start losing your users because it can become too complex for them,” Mr Hayati said.
“We’re looking at the behavioural patterns of spambots and humans such as the way they navigate from page A to page B within a website.
“Currently, spammers just want to submit spam comment and leave, which differs from the pattern of genuine users.”
The most popular topics used in the spam communication analysed in the forthcoming paper were adult themes (about 41 per cent), movies (31 per cent), free offers (15 per cent) and dating (12 per cent).
Kevin Chai, who recently submitted his PhD in web data mining, said that spam busting in an academic environment was not without its challenges.
“Spammers use anti-spam publications to bypass the latest spam filters,” Mr Chai lamented.
“So, in anti-spam we tend not to publish too many of the details.”
Mr Chai said the challenge was to manage spam because the problem would probably never be totally solved.
“It’s like a war,” he said.
“They build a bigger spear and we build a bigger shield and the cycle repeats.”
Mr Chai’s research examines the behaviour of humans in social networks, to provide a benchmark from which to expose the aberrant online behaviour of spammers.
Using the results, the lab may one day be able to eliminate the need for CAPTCHA codes that tend to bug real people wishing to leave a comment on a blog post or online news story.
International estimates put the global cost of spam in the billions of dollars.
The Anti-Spam Research Lab acknowledges contributions of other researchers including Professors Elizabeth Chang (Director of the DEBI Institute), Tharam Dillon, William F. Smyth, Dr Alex Talevski, Faridah Ridzuan, Saeed Sarencheh and Elham Yeganeh.
The Lab will host the eighth annual Collaboration, Electronic Messaging Anti-Abuse and Spam Conference in Perth on September 1 and 2.
Photography: Sam Proctor