There was news from Google that machine learning algorithms were able to block or remove over 170 million policy violating reviews from Google Maps.
Additionally 12 million fake business profiles were removed or blocked.
There were 14 million policy-violating videos detected, while there were over 2 million attempts by bad actors trying to claim Business Profiles that did not belong to them
All this raises a number of questions and comments.
When you see numbers of this size, there can’t be any doubt that whether it’s reviews, profiles or anything else to do with Google Maps, people want to manipulate and distort the truth.
Thus Google has to be on top of their game to catch the offenders.
Using machine learning to help detect the fake ones is a good idea, but the bad actors are probably using machine learning to create them in the first place. Are we entering a machine learning cat and mouse game whereby the good guys are always one step behind?
When there are around 20 million contributions per day on Maps and Search, it’s impossible that humans manually check the veracity of the contributions.
As long as one can submit a review without any form of verification, manipulation is going to happen on an industrial scale. Not only can a fake profile submit a review but they can do so without any proof they were ever a customer in the first place.
Prevention is better than cure and not allowing fake reviews from being posted in the first place is more efficient than removing them after the fact, but it would appear that doing so is simply impossible given the volume of reviews that Google receives.
How would it even work? One has a copy of the receipt, uploads it, it’s verified and then you can post what you’ve got to say?
This is way too much work and so people or bots can post a 1 or 5 star review without writing anything or ever having stepped foot inside the establishment.
Manipulation defeats the purpose of the exercise. Whether it’s good, bad or somewhere inbetween you want to know what it will be like so you can adjust your expectations accordingly. If a place has horrible reviews but you think you will try them out, you may be pleasantly surprised.
170 million policy violating reviews is a frightening number and it’s fantastic that this many were removed, but it makes you wonder how many more fake ones are still live. If Google receives around 20 million contributions a day, this 170 million represents a little bit over a week. What about the contributions from the other 50 weeks of the year!