Eight tech companies pledged to share threat intelligence against online deception on March 16, 2026. Google, Microsoft, Meta, Amazon, OpenAI, Adobe, LinkedIn, and Match Group put their names on the document. The "Online Services Accord Against Scams" was presented at the opening of the first UN Global Fraud Summit in Vienna. Global costs reached $1.03 trillion in 2024, according to GASA.
US consumer losses alone hit $16.6 billion that year. The FBI's Internet Crime Complaint Center recorded 859,532 complaints, a 33% jump over 2023.
We can't solve this alone. We need others across the industry to unite in the effort to tackle scams more collectively.
— Karen Courington, VP of consumer trust experiences, Google trust and safety team
The agreement formalizes cooperation that previously happened ad hoc. Signatories commit to exchanging best practices through GASA and the Tech Against Scams Coalition, and to enhancing user safety features.
Meta's Nathaniel Gleicher told Axios the pact provides a standing venue for sharing information beyond one-off cases. Companies can now compare which defensive measures work and track how threat actors adapt. Microsoft's Steven Masada, assistant general counsel of the Digital Crimes Unit, expects faster communication between partners. He anticipates more disruption operations "designed to be more effective in taking down infrastructure and identifying threat actors."
Deception networks rarely stay on one service. Amazon's Scott Knapp, VP of buyer risk prevention, pointed to the cross-platform nature of modern schemes. A typical ring uses social media for initial contact, messaging apps for grooming, and payment tools for extraction. The pact carries no penalties for non-compliance.
The FBI's IC3 report broke down the $16.6 billion in reported damages. Cyber-enabled crime accounted for 83% of total reported damages ($13.7 billion). Investment crime led at $6.57 billion, driven by cryptocurrency "pig butchering" schemes. People over 60 lost $4.9 billion, a 43% jump year-over-year. The World Economic Forum estimates that only 0.05% of cybercrimes result in prosecution.
Six of the eight signatories are leading AI providers. That dual role matters. Their tools enable criminals to craft localized phishing messages in any language, generate realistic documents, and build convincing personas at scale. Their own products are where much of the resulting activity occurs.
The March 2026 agreement is the third such voluntary pledge in three years. In November 2023, the UK government secured the Online Fraud Charter from 12 companies, including many of the same signatories. No public accountability report on its results has followed. In February 2024, 20 firms endorsed the AI Election Accord at the Munich Security Conference. That pledge expired with the 2024 election cycle.
Google, GASA, and OXIL Research launched the Global Signal Exchange (GSE) in July 2025. The real-time intelligence-sharing tool already includes Microsoft and Meta. The GSE is the operational infrastructure through which the new commitments may flow.
The UN Summit runs March 16-17 at the Vienna International Centre. GASA, Meta, and Santander provided financial support. The Trump administration issued an executive order earlier in March 2026 directing agencies to prioritize cracking down on international criminal rings targeting US consumers.
The agreement formalizes intelligence sharing that was already happening informally. The UK's 2023 Online Fraud Charter used the same voluntary model and published no impact assessment. With $16.6 billion in US damages and a 0.05% prosecution rate, the gap between detection and enforcement is the obstacle that no corporate pledge can bridge alone.
— Artem Safonov, Threat Analyst at AnonHaven
The FBI made 215 arrests in 2024 through 11 joint operations with India's Central Bureau of Investigation. The targeted call centers ran tech support and government impersonation schemes. That enforcement cooperation exists independently of the new pact.
Have a story? Become a contributor.
We work with independent researchers and cybersecurity professionals. Send us a tip or submit your article for editorial review.
