The 3 Biggest Fraud Trends That Emerged at MRC Vegas 2026

At MRC Vegas 2026, gathering 2,000+ payments leaders, the shared theme was that fraud has become more automated and harder to catch with existing tools.

The most advanced response was a strategy of not setting the same barrier for every user, but instead judging trust based on behavior and intent, and applying authentication only when needed. Airbnb's Roberta Del Monte Radford described this as high-trust velocity, saying that friction should be left only for the truly risky 1% of traffic. Stripe's adaptive 3DS also uses AI to assess risk and requests authentication only when something looks suspicious, delivering a fraud reduction of 30% or more on eligible transactions.

The second shift is agentic commerce. Ashley Furniture found it difficult to stop AI agent payments with existing rules-based fraud operations alone, given differing product types and shipping schedules, and emphasized that fraud detection needs to be built directly into the payment fabric rather than applied as after-the-fact analysis.

• Shared Payment Tokens let AI agents pay using stored payment methods without exposing sensitive card information.
• When used together with Stripe Radar, they deliver risk signals—such as the likelihood of a fraudulent dispute, card testing, stolen card use, and issuer declines—in real time.

The third is deepfakes and synthetic identities. H&R Block's Gordon Sheppard demonstrated that with just a single photo, 30 seconds of audio, and about 20 minutes, he was able to create a synthetic video of himself speaking multiple languages. Now, rather than relying on single-point verification, it's necessary to catch subtle anomalies—like signature errors or mirrored facial images—that fraudsters struggle to get perfectly right, across multiple stages.

Stripe Identity programmatically verifies the identities of customers worldwide, detecting fake IDs and spoofed photos, matching ID photos against selfies, and cross-checking SSNs and addresses against global databases for verification.