Try harder

How hard would it be to write some learning algorithms to detect scammers on Linkedin?

Not hard. Their behaviour is quite different to the norm and there is a lot of data to work with.

And then you would just ask the suspects to go through a tougher than normal profile identification program.

There seems to be two types of scams in Linkedin:

1. Totally false people looking to extract something (money) off people. An example was a nice looking young woman that connected with me some time back, with some totally falsified story. Her follow up message had genuine scam written all over it – here is a quote:

“Based on this statistics, Kwesi Bunso family resident has decided to request for genuine foreign partnership with very truthful and honest buyers or investors who will come down to buy their product (Gold Dust or Cooper Ore) and in return, pay for the produce part in cash and the balance”. Really!

I just checked and her profile is still up and running, as thin as it is. She claims education at the business school at Oxford Uni. Her prior work experience is one job as a beauty technician and now she is a director of a gold mine. Heh – I could write an algo to catch this one.

2. The there are real people using Linkedin as a channel for unwanted sales. I had an example the other day where a real estate sales person from Queensland sent me a connection request. I glanced at it and it seemed legit and she (they are often ‘she’s’) had some links in common (more fools). Anyway after we connected I was duly sent a sales pitch:

“Thanks For the Linkedin connection. Glad to connect. Wish to make a positive cash flow property investment in Queensland?”

At least it was upfront. Linkedin should find these guys are force them to go through their own sales and marketing channels in Linkedin so that (a) we can easily identify and ignore them, and (b) Linkedin could get some income off these guys.