In a blog post, Google said it has updated its machine learning systems to detect and remove more fake local business reviews, fake business listings, and fraudulent images and videos. .
Automated systems and human review teams removed more than 200 million photos, 7 million videos and blocked or deleted more than 115 million reviews, representing a 20% increase from the previous year (2021).
How Google Intercepts User-Generated Spam and Detects Fake Local Business Reviews
Google uses brand new machine learning models to identify and remove fake and fraudulent content.
These machine learning models look for unusual patterns in user-contributed content, including detecting new forms of abuse that were previously undetected.
Google announced
“We have a long history of using machine intelligence to identify potential patterns of abuse and continue to develop our technology.
Last year, we extensively overhauled our machine learning models, allowing us to detect new abuse trends several times faster than in previous years.
For example, our automated systems saw a sudden increase in business profiles with sites ending in .design or .top – something that would be difficult to detect manually among millions of profiles.
Our team of analysts quickly confirmed that these sites were fake – and we were able to quickly remove them and deactivate the associated accounts.”
Google’s systems check new content before it is published to block falsified or fraudulent content transmitted to the Google Maps system.
They also use a machine learning model to analyze previously published content and identify fake content that might have missed initial checks. No one really wants fake local business reviews.
These new systems block spam faster than in 2021 and intercept more of it.
Google explains:
“In some places, scammers have started superimposing fake phone numbers onto attached photos, hoping to trick unsuspecting victims into calling the scammer rather than the real company.
To combat this problem, we developed a new machine learning model that can identify overlapping numbers by analyzing specific visual details and photo layout.
Using this model, we were able to identify and block the vast majority of these fraudulent and policy-breaking images before they were published.”
Spam Blocking Statistics
In its announcement, Google specifies that this will be the case in 2022:
- Google has blocked or removed more than 115 million reviews and said most of them were blocked before they were published.
- New anti-spam algorithms removed more than 200 million photos and more than 7 million videos that violated Google’s content guidelines.
- 20 million attempts to create fake business profiles were blocked.
- Enhanced protection for more than 185,000 businesses that have detected suspicious activity.
In January 2023, Google sent a statement to the FTC (read the PDF here), which stated that Google uses signals to detect fake accounts, in addition to content verification.
Google also said it now scans images to detect content embedded in images that is intended to redirect calls from a business to the scammer’s phone number.
They look for bots, duplicate content and word patterns that resemble known fake reviews and also use a system they call “intelligent text comparison” which helps detect misleading content.
Authentic, safe and reliable
Google uses automatic and human reviewers to prevent inauthentic activity in the Google Maps ecosystem. Fake local business reviews harm Google’s reputation and credibility. But also to us, entrepreneurs, if you cannot count on the opinions of your customers.
Detecting fraudulent activity on Google Maps is important both for people who rely on business reviews and for businesses who list their businesses in the system. Internet reviews are important for local marketing.
