Why is Meta Deleting 6.8 Million WhatsApp Scam Accounts?

Meta removes 6.8 million WhatsApp accounts linked to scam operations in the first half of 2025, marking the messaging platform’s largest enforcement action against criminal networks exploiting AI for fraud.

The social media company works with OpenAI to disrupt criminal organisations using AI-generated messages to initiate contact with potential victims.

Many of the deleted accounts connected to organised crime centres in Southeast Asia, where criminals operate under conditions that authorities describe as modern slavery. 

These operations have evolved into billion-dollar enterprises through schemes known as “pig butchering” scams.

According to the US Secret Service, pig butchering scams involve fraudsters gaining victim trust, often through fictitious romantic relationships, before convincing them to invest in fake cryptocurrency projects. 

These schemes typically begin on dating websites or through unsolicited messages on social media platforms.

Inside WhatsApp implementing AI detection systems

WhatsApp has deployed machine learning technologies – computer systems that learn from data patterns – to identify suspicious behaviour before scammers can operationalise accounts. 

The platform now warns users when unknown contacts add them to group chats, a common tactic for fraudsters seeking to expand their reach.

Meta states it “proactively detected and took down accounts before scam centres were able to operationalise them,” indicating a shift toward prevention rather than reactive measures following user reports.

The collaboration with OpenAI targeted a Cambodian criminal group that used ChatGPT to generate messages establishing first contact with victims. 

The sophisticated operation involved AI-crafted text messages containing WhatsApp links, after which targets were moved to Telegram, another messaging platform and assigned tasks like engaging with TikTok videos for monetary rewards.

Scammers would demonstrate fabricated success through screenshots before requesting cryptocurrency deposits to access higher-paying tasks. 

The promise of earnings served as bait in these schemes.

Criminal networks exploiting AI advancement

The enforcement action highlights how criminal organisations have adopted AI tools to scale their operations. 

While Meta and other technology companies deploy AI for fraud detection, criminal networks use the same technology to create convincing communications and automate victim outreach.

Countries including Myanmar, Cambodia and Thailand have become centres for these operations.

Local authorities have struggled so far to contain networks that flourished during the pandemic and have extracted tens of billions of dollars from victims globally.

These operations often rely on forced labour, with individuals recruited under false pretences and compelled to carry out scams. 

The human trafficking element adds complexity to law enforcement efforts.

Consumer rights organisation Which? welcomed Meta’s announcement but called for expanded efforts across all company platforms.

“Meta must do much more to stop these criminals across all its platforms,” the organisation says.

The challenge extends beyond individual companies to regulatory frameworks. 

In the UK, communications regulator Ofcom faces pressure to enforce the Online Safety Act more robustly in response to widespread scamming activities.

Cybercrime has become increasingly difficult to combat due to emerging technologies including AI. 

Criminal organisations adapt their tactics continuously, requiring platforms to anticipate and counter new methods before they achieve scale.

The deletion of 6.8 million accounts sends a clear signal that WhatsApp is serious about doing the dirty work, but, in truth, this victory may prove to be just the beginning of a long road ahead.