Google removed over 5.5 billion ads in 2023, a slight increase from the prior year, according to the company’s annual Ads Safety Report.
The Google Ads Safety report highlighted the tech giant’s goal to maintain a safe and trusted ad experience for users. In 2023, Google nearly doubled the number of suspended advertiser accounts to 12.7 million, compared to the previous year.
This increase shows Google’s growing ability to detect and remove bad actors from its advertising platforms.
AI Plays a Key Role in Ad Safety
This target on malicious ads was driven in part by advancements in artificial intelligence (AI), particularly Large Language Models (LLMs).
The rise of generative AI, including LLMs, is having a huge impact on Google’s ad safety measures.
LLMs are a type of AI that can analyze extensive amounts of information and identify patterns, making them ideal for detecting policy violations in ads.
Traditionally, machine learning models used for ad safety required extensive training on massive datasets. LLMs, however, can learn and adapt more quickly, allowing them to identify new trends in scams and misleading content.
This is particularly beneficial for combating financial scams, where bad actors frequently change tactics to evade detection.
Combating Evolving Threats, From Scams to Deepfakes
In 2023, Google observed a rise in scams across online platforms, prompting the company to update its policies and deploy rapid-response enforcement teams.
One key initiative was the launch of the Limited Ads Serving policy. This policy limits the reach of new advertisers until they establish a commendable record of good behaviour, helping to prevent scams and misleading ads.
Another challenge Google faced was a targeted campaign using deepfakes, which are manipulated videos or audio recordings that appear real.
Google responded swiftly by creating a team to identify and remove these deceptive ads. The company also updated its misrepresentation policy to more effectively suspend the accounts of bad actors using deepfakes.
In total, Google blocked or removed over 206 million ads for violating its misrepresentation policy (which encompasses scams) and 273 million ads for violating its financial services policy. Additionally, over 1 billion ads were blocked for promoting malware.
Investing in Election Integrity and User Control
In the report, Google acknowledged the importance of political ads in democratic elections. The company upholds strict identity verification and transparency requirements for election advertisers, along with limitations on targeting options.
All election ads must include a “paid for by” disclosure and are recorded in a publicly available transparency report. In 2023, Google verified over 5,000 new election advertisers and removed more than 7.3 million election ads from unverified advertisers.
Google remains focused on staying nimble and adapting to new challenges. The company plans to continue developing new policies, strengthening enforcement systems, facilitating industry collaboration, and empowering users, publishers, and advertisers with more control over their online advertising experience.