Researchers have spotted a problem with Google’s robot detecting
YouTube ads might be a great way of getting your business’s message out, but only if real people see them.
Unfortunately for marketers, it has been revealed that Google has been charging companies for ad views, even when its own system detects the video is being watched by a robot and not a real human.
European researchers uploaded videos to YouTube and bought ads on them. They then directed some robots, or “bots”, to the videos and looked at what Google, which owns YouTube, charged them for the ad views.
On one video which had been visited by bots 150 times, YouTube’s view counter showed 25 views, however the researchers had been charged for 91 views.
This means Google had billed the researchers, which were from UC3M, Imdea, NEC Labs Europe and Polito, for views it knew were not by real people.
Assistant professor at UC3M, Rubén Cuevas, said anyone with basic knowledge of coding could have done the experiment.
“It’s not like we found a weird corner-case vulnerability,” he said.
Google said it took “invalid traffic very seriously” and would contact the researchers to discuss the findings.