Today we decided to put together a bunch of facts about cloaking, and although we initially thought we knew almost everything, a few things caught us by surprise:

1. FB dislikes cloaking so much that Meta has filed several lawsuits against different companies over the past few years. The first case was a lawsuit against Basant Gajjar in April 2020 and his software «LeadCloak», which was used to hide websites advertising supplements, cryptocurrencies and so on.

2. The biggest recent incident happened in February 2026, when Meta filed lawsuits against advertisers from Brazil and China for using celebrity deepfakes + against an advertiser from Vietnam for cloaking.

3. The volume of scam ads on FB (which obviously use cloaking) is around $16 billion (10% of Meta's revenue).

4. Platforms prefer to fight cloaking by deploying AI - Google has been the most active, working with Google Research and DeepMind to roll out multimodal LLMs that visit websites and apps and imitate the behavior of real users.

5. About 97% of all Google bans and account suspensions already come from AI detecting patterns (so an account can be restricted before a campaign even launches).

6. By the way, cloaking accounted for 78.2% of Google account suspensions according to the Ads Safety Report 2025.

7. Meanwhile, besides tightening moderation, FB is focusing on increasing the share of ad spend from verified advertisers from 70% to 90%.

8. It's also worth mentioning the rise of AI cloaking. In October 2025, SPLX researchers described «AI-targeted cloaking», where a website serves one version of its content to regular browsers and another to ChatGPT and Perplexity crawlers, telling them apart through a simple user-agent check.

9. In response, agents are starting to use reverse cloaking themselves - for example, Amazon claims that the Comet agent disguised itself as Google Chrome to access a shopper's account.

10. A mechanism for verifying agents (Web Auth Bot) is being developed specifically to stop bots from deceiving websites. In theory, this could make cloaking from such bots much easier in the future.

Today we decided to put together a bunch of facts about cloaking, and although we initially thought we knew almost everything, a few things caught us by surprise:

1. FB dislikes cloaking so much that Meta has filed several lawsuits against different companies over the past few years. The first case was a lawsuit against Basant Gajjar in April 2020 and his software «LeadCloak», which was used to hide websites advertising supplements, cryptocurrencies and so on.

2. The biggest recent incident happened in February 2026, when Meta filed lawsuits against advertisers from Brazil and China for using celebrity deepfakes + against an advertiser from Vietnam for cloaking.

3. The volume of scam ads on FB (which obviously use cloaking) is around $16 billion (10% of Meta's revenue).

4. Platforms prefer to fight cloaking by deploying AI - Google has been the most active, working with Google Research and DeepMind to roll out multimodal LLMs that visit websites and apps and imitate the behavior of real users.

5. About 97% of all Google bans and account suspensions already come from AI detecting patterns (so an account can be restricted before a campaign even launches).

6. By the way, cloaking accounted for 78.2% of Google account suspensions according to the Ads Safety Report 2025.

7. Meanwhile, besides tightening moderation, FB is focusing on increasing the share of ad spend from verified advertisers from 70% to 90%.

8. It's also worth mentioning the rise of AI cloaking. In October 2025, SPLX researchers described «AI-targeted cloaking», where a website serves one version of its content to regular browsers and another to ChatGPT and Perplexity crawlers, telling them apart through a simple user-agent check.

9. In response, agents are starting to use reverse cloaking themselves - for example, Amazon claims that the Comet agent disguised itself as Google Chrome to access a shopper's account.

10. A mechanism for verifying agents (Web Auth Bot) is being developed specifically to stop bots from deceiving websites. In theory, this could make cloaking from such bots much easier in the future.

Today we decided to put together a bunch of facts about cloaking, and although we initially thought we knew almost everything, a few things caught us by surprise:

1. FB dislikes cloaking so much that Meta has filed several lawsuits against different companies over the past few years. The first case was a lawsuit against Basant Gajjar in April 2020 and his software «LeadCloak», which was used to hide websites advertising supplements, cryptocurrencies and so on.

2. The biggest recent incident happened in February 2026, when Meta filed lawsuits against advertisers from Brazil and China for using celebrity deepfakes + against an advertiser from Vietnam for cloaking.

3. The volume of scam ads on FB (which obviously use cloaking) is around $16 billion (10% of Meta's revenue).

4. Platforms prefer to fight cloaking by deploying AI - Google has been the most active, working with Google Research and DeepMind to roll out multimodal LLMs that visit websites and apps and imitate the behavior of real users.

5. About 97% of all Google bans and account suspensions already come from AI detecting patterns (so an account can be restricted before a campaign even launches).

6. By the way, cloaking accounted for 78.2% of Google account suspensions according to the Ads Safety Report 2025.

7. Meanwhile, besides tightening moderation, FB is focusing on increasing the share of ad spend from verified advertisers from 70% to 90%.

8. It's also worth mentioning the rise of AI cloaking. In October 2025, SPLX researchers described «AI-targeted cloaking», where a website serves one version of its content to regular browsers and another to ChatGPT and Perplexity crawlers, telling them apart through a simple user-agent check.

9. In response, agents are starting to use reverse cloaking themselves - for example, Amazon claims that the Comet agent disguised itself as Google Chrome to access a shopper's account.

10. A mechanism for verifying agents (Web Auth Bot) is being developed specifically to stop bots from deceiving websites. In theory, this could make cloaking from such bots much easier in the future.

Today we decided to put together a bunch of facts about cloaking, and although we initially thought we knew almost everything, a few things caught us by surprise:

1. FB dislikes cloaking so much that Meta has filed several lawsuits against different companies over the past few years. The first case was a lawsuit against Basant Gajjar in April 2020 and his software «LeadCloak», which was used to hide websites advertising supplements, cryptocurrencies and so on.

2. The biggest recent incident happened in February 2026, when Meta filed lawsuits against advertisers from Brazil and China for using celebrity deepfakes + against an advertiser from Vietnam for cloaking.

3. The volume of scam ads on FB (which obviously use cloaking) is around $16 billion (10% of Meta's revenue).

4. Platforms prefer to fight cloaking by deploying AI - Google has been the most active, working with Google Research and DeepMind to roll out multimodal LLMs that visit websites and apps and imitate the behavior of real users.

5. About 97% of all Google bans and account suspensions already come from AI detecting patterns (so an account can be restricted before a campaign even launches).

6. By the way, cloaking accounted for 78.2% of Google account suspensions according to the Ads Safety Report 2025.

7. Meanwhile, besides tightening moderation, FB is focusing on increasing the share of ad spend from verified advertisers from 70% to 90%.

8. It's also worth mentioning the rise of AI cloaking. In October 2025, SPLX researchers described «AI-targeted cloaking», where a website serves one version of its content to regular browsers and another to ChatGPT and Perplexity crawlers, telling them apart through a simple user-agent check.

9. In response, agents are starting to use reverse cloaking themselves - for example, Amazon claims that the Comet agent disguised itself as Google Chrome to access a shopper's account.

10. A mechanism for verifying agents (Web Auth Bot) is being developed specifically to stop bots from deceiving websites. In theory, this could make cloaking from such bots much easier in the future.