Google Uses AI To Detect Fraud, Advances Multimodal Search

Google claims it has developed ways through artificial intelligence (AI) to fight fake online reviews, identify fraudsters, and catch scammers. On Monday, it
also highlighted its latest global security and privacy report.

Since Google based the fraud detection on AI, Gemini only needs a few examples to learn what is suspicious, and can apply this knowledge across many business categories and languages.

Some businesses attempt to cheat the system by buying fake five-star reviews from people who have never visited or used the company’s products. AI helps Google to better detect this kind of dishonest behavior.

Through AI, for example, Google can revisit reviews more frequently to identify new abuse patterns even months after they were originally posted.

advertisement

advertisement

Google rolled out alerts in the U.S., U.K. and India to let companies know if suspicious five-star reviews were removed. These warnings, which will expand globally starting in May, could help consumers understand quickly if a place may be engaging in unfair review practices.

AI has been a pivotal tool in helping Google stop scammers. Last year, Google removed more than 10,000 listings managed by a group of bad actors who impersonated real locksmiths to take over unclaimed Business Profiles and overcharge unsuspecting customers.

Google, for example, trained a new model with the help of Gemini that identifies potentially suspicious profile edits. A business that changes its name from “Zoe’s Coffee House” to “Zoe’s Cafe” isn’t suspicious — but the tech watches and identifies business that suddenly changes its category from “cafe” to “plumber.”

Google said that based on AI in 2024, it blocked or removed more than 240 million policy-violating reviews from 2024, the vast majority of which were removed before they were seen.

Improved AI detection blocked or removed more than 70 million policy-violating edits to places on Maps. Google also removed or blocked more than 12 million fake Business Profiles.

Google placed posting restrictions on more than 900,000 accounts that repeatedly violated its policies.

The company today also began to make AI Mode available to millions more U.S. Labs users, as well as brought multimodal capabilities in Lens to AI Mode.

On average, AI Mode queries are twice as long as traditional Search queries on Google. More words in a query helps to focus in on the response. People also use AI model for help to explor open-ended questions and complicated tasks like comparing two products, exploring how-tos and planning a trip. 

Next story loading loading..