Phoenix 24
A global digital strike targets criminal networks while boosting real-time user protection through AI-powered detection.
New York, August 2025.
In one of the most extensive fraud-prevention operations in its history, WhatsApp has deleted over 6.8 million user accounts linked to coordinated scams during the first half of the year. Meta, the company behind the messaging giant, described the action as a strategic response to the rise of global cybercrime targeting users across borders, platforms, and digital habits.
According to internal reports, many of these accounts originated from Southeast Asia, particularly Cambodia, where large-scale scam centers were operating schemes ranging from fake cryptocurrency investments to fraudulent job offers and social media “engagement” rewards. These operations typically exploited emotional manipulation, financial desperation, or digital naivety to extract personal information and money from victims.
What made this takedown different was the technical collaboration involved. Meta confirmed joint efforts with OpenAI to dismantle a criminal network that was using generative AI tools—such as ChatGPT—to craft highly personalized and persuasive scam messages. These messages were deployed via WhatsApp and then redirected victims to alternate channels like Telegram or TikTok, deliberately evading detection systems.
In response, WhatsApp has rolled out new security features aimed at disrupting scams before they unfold. These include a “security context preview” when users join groups created by unfamiliar contacts, and proactive alerts in one-on-one chats that signal potentially risky interactions before they escalate.
The broader context is alarming. According to the U.S. Federal Trade Commission, online scams cost consumers over $12.5 billion in 2024—an increase of nearly 25% from the previous year. Experts point out that fraud has evolved beyond crude impersonation: scams are now emotionally strategic, technologically refined, and socially engineered to feel legitimate.
El mensaje inicial de la estafa era creado en ChatGPT, según OpenAI. REUTERS/Dado Ruvic/Ilustración
Meta’s internal detection mechanisms played a key role in this crackdown. By analyzing anomalous behavior patterns and engagement signals, the system blocked many accounts before their scam activity had even begun. This reflects a growing trend in cybersecurity: prevention through predictive AI rather than just post-incident response.
The operation also reveals a shift toward collaborative digital policing. By combining Meta’s internal monitoring systems with OpenAI’s linguistic and behavioral analysis capabilities, the initiative neutralized fraud networks that span continents and platforms—suggesting that fighting cybercrime now demands multi-layered, multi-platform coordination.
While encryption remains a cornerstone of WhatsApp’s promise to its users, this case underscores the paradox facing all secure messaging services: maintaining privacy while protecting against abuse. By embedding real-time safety cues and expanding its detection protocols, WhatsApp is signaling a move toward active defense—neutralizing threats before they reach the end user.
In the words of a cybersecurity analyst consulted by Phoenix24, “Scammers no longer work in isolation. They function like startups—fast, iterative, data-driven. The only way to beat that is with systems that are equally agile and globally integrated.”
“This piece was developed by the Phoenix24 editorial team using reliable sources, public data, and rigorous analysis in alignment with the current global context.”
“Esta pieza fue desarrollada por el equipo editorial de Phoenix24 con base en fuentes confiables, datos públicos y análisis riguroso, en coherencia con el contexto global vigente.”