Invention of the day 7

An algorithm designed to detect a real person from a scamming identity on Linkedin.

You can think of this problem as the Reverse Turing Test – how a computer can detect a real person from a fake person.

If I can spot a scammer from 100 paces, with my hands tied behind my back, at sunset, with a bag over my head, then it wouldn’t be that hard to put an algorithm together. The basis rule of thumb of data and image processing algorithms is that if the human brain can spot it then the algorithm is possible.

The real question is why Linkedin hasn’t commissioned such an algorithm already.

Almost weekly now I am contacted by some dodgy scammer, usually a blonde woman with a barely plausible resume, and for no apparent upfront reason.

They could run such an algorithm over all new users, or even over the current database, and weed out the fakes.

Or at the very least they could offer it as a plugin for those that don’t want to be scammed by blonde imponderables.

The trouble of course is that Linkedin doesn’t have competition and hence aren’t really on their game.

On a final philosophical note; if a computer can tell a real human from a fake human is that a positive test for artificial intelligence?

I would say yes given that there are plenty of blokes that are supposedly intelligent that actually fall for these blonde facsimiles.

But then the counter argument might be that they do this knowingly at some deep subconscious level, trading off truth for hope.

Such an argument, when boiled down, would suggest that artificial intelligence will only occur when computers have a desire to mate.