Deepfakes and AI scams are booming. Here’s how Meta, banks and police are teaming up to stop them

THERE is a new generation of scams that is becoming increasingly difficult to detect. They can range from a deepfake video of a trusted public figure endorsing an investment to a voice note that mimics a bank employee or a job offer that references your real work history from a credible-looking recruiter. 

Generative AI has lowered the barrier to entry for scam actors while increasing the quality, sophistication, and scale of their operations. Scammers can use AI to create convincing fake accounts with coherent backstories and authentic-looking profile pictures and videos. 

There are also cases where AI facilitates rapid creation of fake businesses – complete with websites, email trails and fraudulent documentation – to scam individuals.

Fighting an international issue

To combat these advances in malicious use of AI, Meta is also using AI to better detect scam activity and protect users across its platforms.

Earlier this year, the company disrupted a Cameroon-based network consisting of 100 groups and approximately 12,000 accounts across Facebook and Instagram that were using AI to facilitate animal rehoming scams. The network posted fake pet adoption content to lure targets, then requested upfront payments under the guise of covering veterinary care, transportation or rehoming fees.

Meta’s investigation revealed that these scammers used AI to scale their operations, generating location-specific content and culturally appropriate messaging to improve targeting in the US, Canada, Australia, New Zealand and the UK.

The company has also removed, disabled and unpublished over 15,000 assets on Facebook and Instagram that used deceptive personas claiming to be Japanese women seeking relationships with millennials and older men, with some accounts also promoting gambling-related content. After luring in users, these scammers would direct conversations to private channels like Line or Messenger.

Meta’s investigation revealed that the operations had a wide reach. They were based primarily in China, Myanmar, Colombia, the Philippines, Indonesia, the US and Nigeria, and they targeted users in Japan and other markets.

These examples are not outliers. Criminal scam syndicates now operate like profitable corporations, complete with recruitment pipelines, performance metrics, and research and development departments.

AI as a sword and a shield

The same technology weaponised by scammers can also be used to stop them.

Meta has built AI systems that analyse multiple signals – like text, images, and surrounding context – to spot sophisticated scam patterns that traditional detection would miss.

These models can identify when a scammer is impersonating a public figure or brand by analysing fake fan sentiment, misleading bios, and deceptive associations – the kind of contextual clues that no human team could process at the speed and scale required to keep most users safe. 

The systems can also detect content that redirects users to webpages designed to mimic legitimate ones, catching domain impersonation before people are exposed. However, even with these efforts, technology alone cannot solve a problem rooted in transnational organised crime. 

Typically speaking, before a scam ever reaches a platform, criminals have already registered domains, acquired SIM cards, or built physical compounds. In some cases, scammers use a fake banking app, a cryptocurrency wallet, and a mule account in one country and a call center in another. No single institution can see a scam from end to end. When broken down, it looks like this:

• Platforms see the fraudulent post.
• Banks see the suspicious transaction.
• Police see the victim report.

Scammers work by exploiting the gaps between these institutions. Closing those gaps requires intelligence to flow across sectors in real time, and across Asia Pacific, new models of cross-sector collaboration are emerging to address those issues.

The critical role of partnership and intelligence sharing

In Sept 2025, the Government Technology Agency of Singapore (GovTech Singapore) became the first government agency in the world to join the Global Signals Exchange (GSE), a centralised threat-intelligence organisation that enables real-time sharing of scam intelligence between government, platforms and the financial sector. 

This has already led to a significant impact: Between Oct 2025 and Feb 2026, Meta removed over 30,800 entities and pages associated with fraud and scams from Facebook and Instagram based on URL signals shared by GovTech Singapore through the GSE. The two organisations have since implemented an automated reporting system, allowing threats to be flagged and acted on at machine speed rather than human speed.

This kind of signal sharing also powers enforcement on the ground. Since Dec 2025, Meta has worked with the Royal Thai Police and FBI to host three Joint Disruption Weeks, bringing together law enforcement agencies from around the world to participate in live signal-sharing exercises. 

In May 2026, Meta, Microsoft, Coinbase, Starlink and other leading tech companies teamed up with the US Department of Justice, the Royal Thai Police, and global law enforcement agencies – including law enforcement partners in Australia, Canada, Indonesia, Japan and Malaysia, among others – to share information and disrupt criminal scam networks in joint operations across the US and Thailand. 

More than a million online assets were disrupted as a result of the operation, including 1.4 million accounts, pages, and groups across Facebook and Instagram; 20,000 Microsoft accounts; and thousands of Starlink kits. The Royal Thai Police also arrested 63 individuals involved in scam operations.  

These efforts are also informing broader multilateral policy development. At the 2026 APEC Digital Weeks in Chengdu, China, Meta worked with APEC country representatives and tech companies to examine the role of public-private partnerships in strengthening scam and fraud prevention efforts.

The discussions acknowledged the importance of a whole-of-ecosystem approach and outlined practical pathways for deeper public-private collaboration

These efforts are helping cross-sector stakeholders make meaningful progress in disrupting online criminal scam networks and protecting internet users around the world.

The technology to detect and disrupt scams at scale exists, and the partnerships to act on shared intelligence are proving their value. What is needed now is the commitment – from platforms, governments, financial institutions and law enforcement alike – to make this level of coordination permanent, not episodic.

Scams affect people across platforms and industries. Meta takes a comprehensive approach to making its platforms safer. Learn how it is working to help keep users safer on its technologies. This article, first published in Tech in Asia, is written by Beth Ann Lim, director of Public Policy Strategy APAC at Meta.