trending
AU10TIX has issued a stark warning to the global payments and digital identity ecosystem: fraud is no longer episodic, reactive or human-led.
The future of identity fraud
According to its Signals for 2026 special edition, the industry has entered the “Year of Machine Deception”, a turning point in which fraud systems learn, adapt and redeploy autonomously.
The report argues that identity fraud has evolved from isolated attacks into interconnected, self-optimising ecosystems.
What once appeared as repetitive, low-level fraud patterns — so-called “repeaters” — are now understood as training data for adaptive fraud engines operating across platforms, regions and sectors.
From Discrete Attacks to Living Systems
At the heart of AU10TIX’s analysis is the rise of agentic AI: self-directed systems capable of generating, testing and refining deception without human intervention.
These systems operate via automated feedback loops, learning from both successful and failed attacks to improve future outcomes.
As a result, fraud is no longer best understood as a collection of anomalies. Instead, it behaves like a living signal, propagating through devices, networks and identities.
Traditional detection models, which rely on static thresholds and one-off irregularities, are increasingly blind to this form of systemic deception.
Why 2026 Raises the Stakes Further
Looking ahead, the report highlights two converging risks.
First, agentic AI fraud engines capable of autonomously creating and managing synthetic identities at scale.
Second, quantum risk, which threatens to undermine the cryptographic foundations on which digital trust currently rests.
To address this, AU10TIX has introduced a Predictive Resilience Framework that unifies anomaly intelligence with quantum-resilient cryptography.
By combining hashing, post-quantum-aligned encryption and predictive analytics, the model protects both the data itself and the behavioural signals surrounding it.
Leading Indicators Payments Firms Cannot Ignore
The report forecasts sharp increases in key fraud vectors during 2026, including a 100 per cent rise in presentation spoofing, a 60.7 per cent increase in identity drift, and a 36.4 per cent jump in credential replay attacks.
As Yair Tal notes, anticipating fraud now matters more than reacting to it. In an era where deception learns as fast as defence, the battle for trust is shifting decisively from detection to prediction.
Comments