Spam-trap Turing tests train smarter software

Paul Raven @ 15-10-2008

Email and comment spam is one of those constant low-grade annoyances that simply becomes part of the furniture if you spend a lot of time on the ‘net, as are the CAPTCHA puzzles you have to take to prove you’re a human. [image from Wikimedia Commons]

Signs are that they won’t be much use much longer, though; a UK researcher has been using the ‘twisted letters’ type of CAPTCHA to train his visual recognition algorithms, while a chap at Palo Alto has a program that can correctly identify cats and dogs 83% of the time – which, lets face it, is probably a better success rate than the average YouTube user can manage.

Sadly, training algorithms against Turing test spam-traps is no more likely to produce a recognisably intelligent piece of software than the Loebner Artificial Intelligence Prize is. But maybe one day we’ll be able to combine all the pieces… if they don’t beat us to it and combine themselves, of course. 😉


When 3D spam got old

C Sven Johnson @ 30-09-2008

The latest instalment of Sven Johnson’s Future Imperfect is part of the Superstruct project.

Future Imperfect - Sven Johnson

Within it, a future iteration of Sven takes a moment on a cruise-gone-wrong to reflect on the history of 3D spam – the flipside of the fabrication revolution. Continue reading “When 3D spam got old”


Storm botnet turns its hand to writing fiction

Paul Raven @ 27-06-2008

lightning strikeHere’s a new twist in the ongoing saga of the Storm worm spam network – it has started delivering fiction into our inboxes. [via Bruce Sterling]

Not science fiction, sadly – that’d make for an even better headline – but fake news headlines. Perhaps in response to people slowly wising up to email subject-lines about fake Rolex watches and “spec14l blu3 p1ll 4 b3dr00m”, the botnet is now replacing them with specious news stories about non-existent natural disasters and celebrity mishaps:

“The emails contain such headlines as ‘Eiffel Tower damaged by massive earthquake’ and ‘Donald Trump missing, feared kidnapped.'”

Pitching for the schadenfreude market, then … we’ll be able to judge the effectiveness of this new tactic by watching for how long they keep using it. [image by El Garza]


Psychology researchers inadvertently enable Second Life spam-bots?

Paul Raven @ 06-11-2007

giant laptop in Second Life A group of UK based psychology researchers were interested in seeing how Second Life users reacted to invasions of personal space within the virtual world. So, they developed a way around the built-in limitations that Linden Lab put in place to prevent software-controlled avatars being deployed, enabling them to send an avatar on autopilot to interact with other residents and record their reactions.

To which your response might be "so what?" – especially if you’re skeptical about Second Life to start with, which is not an uncommon stance. But as the heads-up on SlashDot points out, what can be done by psychology researchers in the name of science could just as easily be done by spammers seeking a automated method of advertising in the metaverse … which would seem to reinforce the adage that no platform will ever remain completely immune to spam techniques. Still, at least in SL you can always teleport away from an annoying avatar, which is more than you can do when confronted by a Scientologist or insurance hawker in the high street … [Image by PsychoAl]

[tags]metaverse, Second Life, spam, software, psychology[/tags]

Spammers Defeat CAPTCHA?

Jeremiah Tolbert @ 10-07-2007

In the war between spammers and everyone else, the spammers may have captured new territory. A new trojan appears to be capable of bypassing the CAPTCHA systems on Yahoo and Hotmail, allowing spammers to create 500 bogus email addresses per hour. CAPTCHA tests are the distorted images of text that computers have previously been unable to read. They’re a kind of simple Turing Test meant to require a human behind a keyboard when creating a new email address.

I am suspicious of the claim that the trojan is actually somehow able to read these images, which have thus far been impossible to crack as a security measure. New Scientist Blog agrees. 500 an hour is not very fast. There is some trickery at work here, perhaps in the form of passing the CAPTCHAs from Hotmail to another website where humans are doing the solving work for the spammers.


« Previous Page