Google Ads annual safety report reveals year-on-year rise in scams and fraud; growth in generative AI enforcement against malicious advertising
Google Ads have released their 2023 Safety Report and it makes for interesting reading. Well, it makes for reading. The tl;dr headline takeaways are all in the areas you’d expect:
- Generative AI represents an growing threat to user security
- It’s an even more valuable tool in the mega-corp’s enforcement toolkit
- Fraud and scams are on the rise
- Google’s Limited Ads Serving policy, while still in its infancy, should limit bad actors by reducing the reach of less familiar advertisers
- In 2023 Google launched the Ads Transparency Center, “a searchable hub of all ads from verified advertisers, which helps people quickly and easily learn more about the ads they see on Search, YouTube and Display”
- A dedicated team has been created to combat deepfakes
- They’re also investing heavily in election integrity
This last one, while perhaps not as sexy (to some) as all the AI stuff, is surely the most important of all. This year is set to be the most election-y in history and there is increasing scepticism that the big tech firms are doing their bit to preserve democratic integrity. Google says it is demanding more identity information from advertisers seeking to enter the political space, especially with regards to where the money to pay for them is coming from, and has been taking action action ads that make “demonstrably false election claims.”
2023 in numbers
- 5.5 billion – ads blocked or removed in total
- 1 billion – of these for abusing the ad network, which includes promoting malware.
- 206.5 million – ads blocked or removed for violating Google’s misrepresentation policy
- 12.7 million – advertiser accounts deleted (twice as many as the previous year)
- 2.1 billion – publisher pages on which ads were restricted from appearing
- 395,000 – publisher sites against which Google took “broader site-level enforcement action”
- 90% – proportion of this action that started with machine learning models, including LLMs
- 5,000 – new election advertisers verified
- 7.3 million – election ads removed
This certainly all sounds very impressive and reassuring but what’s missing is much in the way of context.
Is 5.5 billion ads a lot? A billion is a very big number, that’s something on which we can surely all agree, but is it enough, in the scheme of things, for a (part of a) company that generated $237.855 billion in the same year?
The answer to that can be found in the question. As long as advertisers keep advertising, and users keep searching, Google will feel the action its taken is sufficient. In short: vote with your feet. (Except on election day, when you should probably mark your cross with your left or right hand.
