Google has used artificial intelligence to suspend more than 39 million accounts of fraud

Google said on Wednesday that she had suspended 39.2 million accounts announced on her platform in 2024 – more than three times the number from the previous year – in the latest campaign to fraud ads.

By taking advantage of the LLMS models and using signals such as impersonating the business person and illegal payment details, the research giant said it could suspend a “vast majority” of advertising accounts before submitting an advertisement.

Last year, Google launched more than 50 LLM to improve safety enforcement mechanisms on all its platforms.

“While artificial intelligence models are very important to us and have provided a series of impressive improvements, we still have human beings during the process,” said Alex Rodriguez, the Google Advertising Secretary General, said.

The executive director of the journalists told a team of more than 100 assembly experts throughout Google, including members of the advertising safety team, the confidence and safety department, and DeepMind researchers. They analyzed Deepfake’s reports that involve public personal suicides and develop counter measures.

The company presented the technical measures and more than 30 advertisements and invited the publisher’s policy last year. These moves helped to suspend more than 700,000 offensive advertising accounts, which led to a 90 % decrease in DeepFaken reports, as the company claims.

In the United States alone, Google said it had stopped 39.2 million announced accounts and reduced 1.8 billion ads last year, with major violations related to the abuse of the advertising network, misuse of trademarks, health care demands, personal ads, and distortion.

Google said that India, the most populated country in the world and the second largest Internet market after China in terms of users, witnessed 2.9 million commentators at the expense of last year, making it the second higher after the United States, and the company also removed 247.4 million ads in India, with the five best violations of policy related to financial services, poor use of trading, abuse of advertising network, personal games, and games.

Among all the suspension of the advertiser’s account, Google said it had suspended 5 million accounts on fraud violations.

In general, the company has removed nearly half a billion declarations related to fraud.

Google also achieved more than 8,900 new advertisers in the elections in 2024, which saw that half of the world’s population goes to the polls, and removed 10.7 million electoral advertisements. However, Rodriguez indicated that the volume of election ads compared to the total ads numbers of Google was relatively small and will not significantly affect safety standards this year.

In total, Google said it had banned 5.1 billion ads last year and removed 1.3 billion pages. In comparison, it is Prohibited More than 5.5 billion ads and took measures for 2.1 billion pages published in 2023.

Google Techcrunch told the decreasing numbers indicated improvements in prevention efforts. By improving early detection and having harmful calculations, fewer ads are produced or access to the platform.

She said that the company restricted 9.1 billion announcements last year.

More importantly, the suspension sometimes raises concerns about the extent to which the company applies its rules. Google said it provides an appeal that includes human reviews to ensure that it takes the “appropriate action.”

“Often, some of our messages were not clear and transparent about the details, or about what was the logical basis, or logic, and sometimes the advertiser left a little more mixed. We ended up updating a set of our policies as it is related to that, as there was a large group of our transparency capabilities in terms of 2024. Rodriguez said.