11 Tech Giants Sign Anti-Scam Accord at UN Summit
Adobe, Amazon, Google, Meta, Microsoft, OpenAI, and six other tech companies signed a voluntary anti-fraud accord at the UNODC Global Fraud Summit in Vienna on March 17, 2026.
6 min read
Eleven of the world's largest technology companies signed a joint commitment to fight online fraud on Monday at the UNODC Global Fraud Summit in Vienna, formalizing pledges that critics have already noted carry no enforcement teeth.
TL;DR
- 11 signatories: Adobe, Amazon, Google, LinkedIn, Levi Strauss, Match Group, Meta, Microsoft, OpenAI, Pinterest, and Target
- Commitments include fraud detection tools, financial transaction verification, and intelligence sharing with law enforcement
- The accord is completely voluntary - there are no penalties for non-compliance
- Apple didn't sign
- Summit co-organized by UNODC and INTERPOL, attended by 1,300+ participants from government, tech, and finance
The accord, formally titled the Industry Accord Against Online Scams and Fraud, was announced at the Vienna International Centre on March 17, 2026, the final day of the two-day summit. It brings together an unusually broad coalition - AI labs, social platforms, an apparel retailer, and a dating company - under a shared recognition that AI has made fraud dramatically cheaper and more effective to run at scale.
"Scammers are growing more sophisticated, and they exploit gaps between companies and platforms - which is why tackling global online fraud demands a united, collective response."
- Nathaniel Gleicher, Meta's Global Head of Counter Fraud
What the Companies Actually Committed To
The accord is a voluntary framework. Signatories agreed to deploy fraud detection tools, implement stronger verification requirements for financial transactions, share threat intelligence with law enforcement, and apply best practices for scam prevention and reporting. They also formally requested that governments "declare scam prevention a national priority."
None of those commitments come with deadlines, metrics, or audit rights. A company can sign the accord and continue operating exactly as it did before, as long as it can point to some fraud-detection work already in progress.
The Companies
Meta's participation carries the most operational weight. The company disclosed that it removed 159 million fraudulent ads in 2025, disabled 10.9 million accounts linked to criminal scam networks, and collaborated with the Royal Thai Police to disable more than 150,000 additional accounts linked to a single organized crime operation, resulting in 21 arrests. Those figures suggest a genuine enforcement apparatus, though they also reflect how large the problem is relative to what any one platform can address alone.
OpenAI's presence at the signing is notable. The company published a misuse report on ChatGPT and scam operations earlier this year that documented how state-sponsored actors and cybercrime groups were using the platform. Signing the accord doesn't resolve the tension at the core of that report: OpenAI's own models are among the tools being used to produce the scam content this accord targets.
LinkedIn added recruiter and executive verification requirements as a recent anti-fraud measure. Match Group - owner of Tinder, Hinge, and OkCupid - faces persistent romance scam exposure on its platforms. Target and Levi Strauss represent retail sector exposure to refund fraud and account takeover attacks.
The UNODC Global Fraud Summit brought together over 1,300 participants from government, law enforcement, and the private sector at the Vienna International Centre.
Source: unodc.org
The Users
Consumers should expect slightly better fraud detection on platforms that sign the accord, though the real variable is enforcement effort, not the accord itself. The intelligence-sharing commitment is the provision most likely to produce concrete outcomes: when one platform detects a coordinated fraud campaign, sharing indicators of compromise with other signatories can disrupt operations across multiple services simultaneously.
The financial transaction verification commitment could produce friction for legitimate users. Stronger identity checks at payment points tend to catch fraudsters but also slow down ordinary customers, especially older users or those in markets with weaker digital identity infrastructure.
The Absent Name
Apple didn't sign. The company declined to comment. Its absence is significant given that Apple Pay, the App Store, and iMessage are all vectors for the fraud types the accord covers. Romance scams that begin on third-party platforms often migrate to iMessage to evade platform detection. App Store distribution has been used repeatedly to push fraudulent investment and crypto apps that only reveal their true purpose after installation.
Apple's non-participation doesn't prevent the accord from functioning, but it leaves a visible gap in the coalition's coverage.
Eleven companies at one table sounds like progress. The question is whether a voluntary pledge does anything the existing economics of fraud prevention didn't already demand.
Who Benefits, Who Pays
The accord is structured so that compliance costs fall primarily on companies that haven't already invested heavily in fraud infrastructure. For Meta, Google, and Microsoft, the commitments map closely onto work already underway - Meta's 159 million fraudulent ad removals happened in 2025, before the accord existed. Signing costs them almost nothing.
For smaller signatories with less mature fraud operations, the information-sharing provisions carry real value. Access to threat intelligence from Meta or Google's detection systems is worth more than anything a company like Pinterest or Match Group could produce internally.
| Stakeholder | Impact | Timeline |
|---|---|---|
| Large platforms (Meta, Google, Microsoft) | Minimal operational change - existing programs already exceed accord commitments | Immediate |
| Mid-tier signatories (LinkedIn, Pinterest, Match) | Intelligence-sharing access improves detection; some verification investment required | 6-12 months |
| Retail signatories (Target, Levi Strauss) | Gradual fraud tool deployment required | 6-18 months |
| Users on signatory platforms | Slightly better fraud detection; possible payment verification friction | Gradual |
| Users on non-signatory platforms (Apple) | No direct change from accord | N/A |
| Law enforcement | Access to coordinated threat feeds from 11 platforms simultaneously | As frameworks are established |
Meta detailed its anti-scam technology stack at the summit, including AI tools for detecting coordinated fraud operations across Facebook, Messenger, and WhatsApp.
Source: about.fb.com
What Happens Next
The accord has no secretariat, no annual review process, and no mechanism for removing signatories who fail to follow through. Whether it produces measurable results depends almost completely on whether the companies involved treat it as a binding operational commitment or as a press release.
The UNODC is separately negotiating what it calls a "UN Global Public-Private Partnership Agreement against Fraud" - a more formal framework that Meta has already endorsed. If that agreement advances, it would give the voluntary commitments made in Vienna a more durable institutional home.
The AI dimension of fraud is accelerating faster than any voluntary accord can track. Voice cloning tools - many of them freely available or built on models like those reviewed in our AI voice generator roundup - have made phone-based impersonation scams dramatically more convincing. The AI voice leaderboard tracks how quickly synthesis quality is improving. That improvement cycle runs independently of what any platform decides to do about it.
The summit attracted 1,300 participants from government, law enforcement, tech companies, financial institutions, and civil society. Governments from Canada, Singapore, and the United Kingdom co-sponsored the event. The US did not.
Sources: Meta Newsroom - Fighting Scammers · UNODC Global Fraud Summit · Dataconomy - Major Tech Firms Sign Accord
