According to The Verge,
the AdWords program enables Google to bring billions of targeted ads to
its users. Although the majority of these ads are genuine, some are
linked to phishing schemes and counterfeit items. Much like how an
email’s spam filter operates, Google employs a computer algorithm to
identify any ads that display shady behavior. However, when the program
flagged seemingly harmless used-car ads as suspicious, the team believed
a mistake was made.
Upon the initial flagging, the only logical explanation was that the
vehicles were fake. Dismissing this idea, engineering director David
Baker, said, “We’d never heard of a counterfeit car”. After digging a
little deeper, the team uncovered a popular Chinese car scam. The
criminals were taking pictures of regular cars parked on the street, and
putting them up for sale online. When the customer went to make the
payment a few days later, the scammers would break-in and steal the car,
offering it up to the unknowing victim.
Although it might seem like common sense to request documentation such
as the official car title and registration beforehand, the schemers are
adept at talking their way out of these situations. “These people are
very professional,” added Dahui Li, an information systems expert at the
University of Minnesota.
So how did the computer spot the fraud before the experts? According to
Baker, “there’s no one thing or even a handful of things.” The program
aggregates thousands of pieces of information, such as the poster’s IP
address, user age, and links to past accounts. Perhaps the biggest
tipoff was that the high-priced transactions were occurring on
relatively new accounts.
That being said, to the average person the car ads were completely
innocuous, and it’s baffling that the Google software could identify
such a big crime, especially one that largely took place offline.
No comments:
Post a Comment