Google 2024 Ads Safety Report: AI plays key role in safer advertising ecosystem

The 2024 report focuses on the critical role of AI in enforcing policies at scale, the preventative measures Google is taking to quickly identify and block suspicious ads and advertisers, and the fight against scam ads, particularly deepfake scams. Google’s team of over 100 experts permanently suspended over 700,000 bad advertiser accounts last year, resulting in a significant 90% drop in reports of scam ads. In the run-up to global elections, Google is also taking additional steps to combat ad manipulation.

Some key figures from the report:

• Over 5.1 billion ads blocked or removed for regulatory violations.
• More than 9.1 billion ads were restricted.
• 39.2 million advertiser accounts suspended for serious violations.
• Removed ads from over 1.3 billion web pages.
• Over 8,900 new election ads verified and 10.7 million election ads from unverified accounts removed.
• Over 222,000 ads removed from publisher websites.

LLM’s

Google says it is investing heavily in improving its Large Language Models (LLMs), building on Gemini with more than 50 model improvements. These efforts will help enable more efficient and accurate enforcement at scale. By 2023, more than 97% of Google’s page-level enforcement will begin using machine learning models, including the latest LLMs.

Google’s continued advancements in AI and policy enforcement are seen as essential to protecting the digital advertising landscape. More details can be found in Google’s blog and the full 2024 Ads Safety Report.