Deepfake fraud has surged more than 2,000% in just three years. Criminal networks now operate billion-dollar fraud ecosystems using AI, shell companies and cryptocurrency, moving faster than fragmented enforcement systems can respond. Fighting back requires three fundamental shifts. Luckily, the technology is there, but the real question is whether collaboration can catch up with innovation before public trust collapses entirely.

Financial fraud has evolved into an industrialized, borderless enterprise rivalling global narcotics and human trafficking. Fuelled by artificial intelligence, today’s scams are faster, more convincing and more scalable than ever. Deepfake fraud attempts have surged 2,137% in just three years, AI-generated fake IDs are openly sold on the dark web, employees get tricked into transferring millions during a video call with a deepfake executive… These events are no longer rare, they’re part of a growing $40B global threat projected by 2027.

“At IDnow, we are fully aware of the growing threat posed by social engineering scams,” says Michal Kalinowski, Team Lead Fraud Prevention at IDnow. “These are sophisticated attacks where individuals are psychologically manipulated into opening accounts and authorising fraudulent transactions under false pretenses. Often, victims are lured by promises of high returns, fake job offers, romance scams or unrealistic financial incentives.” In these scenarios, the victims themselves are not committing fraud; rather, they are being exploited as conduits to move stolen money.

Yet the global response remains fragmented and slowed by legacy systems, jurisdictional divides and uneven digital readiness. The question is no longer what needs to be done, but whether we can move fast enough to do it.

“Beyond the traditional social engineering tactics, an even more complex threat is emerging, the financial agent or first-party fraud,” Kalinowski explains. “These are scenarios where users willingly cooperate with criminals to open accounts using their own documents, but with fraudulent intent of selling access to those accounts or directly facilitating money laundering. Here, there is no psychological manipulation or victimhood; there is active, deliberate collusion. Detecting and preventing this intentional type of abuse requires a totally different set of indicators and defense strategies.”

The next phase depends on how effectively regulators, banks and technology providers can align under a shared mission: building a real-time, cross-border defense infrastructure. Europe’s regulatory momentum – from AMLR to eIDAS 2.0 – creates a unique window of opportunity. But unless collaboration catches up with innovation, financial systems risk losing public trust altogether.

The path forward: Collaboration or failure

The financial toll industrialized fraud creates is immense. Across Europe, detected fraud cases have surged 43% year-over-year, from 3.89 to 5.57 per 100,000 transactions. Global fraud losses surpassed $1 trillion in 2024.

For financial institutions, the costs are huge (£21,400 per hour fighting financial crime in the UK alone), but for customers, losses are often life-changing. Many lose their entire savings – only around 4% ever recover their money – and suffer lasting mental health effects.

When people lose confidence that banks can protect their money, that governments can prosecute criminals or that legitimate investments are distinguishable from scams, entire systems begin to fracture.

To reverse the trend, three things must happen – simultaneously and at scale.

1. Triangulation between public, private, and technology sectors

Law enforcement has authority but lacks real-time transaction visibility. Banks have data but not investigative powers. Technology providers have tools but need scale and oversight. Criminal networks, meanwhile, operate with perfect coordination across borders.

Only by breaking down silos and enabling secure, compliant intelligence sharing can we match their speed. Promising frameworks are emerging:

Europol’s European Financial and Economic Crime Centre (EFECC) connects law enforcement, financial institutions and private partners for structured intelligence-sharing and faster cross-border cooperation.
The UK’s Joint Money Laundering Intelligence Taskforce (JMLIT) and the Netherlands’ FIU-NL collaboration model enable near-real-time intelligence exchange between banks, regulators and investigators.
AML Innovation Hubs (like BIS’s Project Aurora) test privacy-enhancing technologies that enable collaborative analysis and learning across institutions and borders, allowing banks and fintechs to detect money laundering networks more effectively while protecting sensitive information.
AI-driven analytics platforms are being embedded into government and financial data infrastructures, enabling pattern detection and alert systems that operate across institutions.

At IDnow, we see collaboration not just as a value but as a core driver of effective fraud prevention. “In a rapidly evolving digital landscape, IDnow must step up the fight against fraud with a shared global fraud intelligence database,” says Kalinowski. “This system leverages high-risk indicators, including geolocation, IP addresses, and user identification history to provide real-time, intelligent risk assessment. This data will be seamlessly shared across our partner network to ensure a unified and proactive defense.”

AI-driven analytics platforms are being embedded into government and financial data infrastructures, enabling pattern detection and alert systems that operate across institutions.

2. Cross-border legal frameworks

Today, fraud operates without borders, but enforcement still does. Europe’s AML landscape remains fragmented across 27 national frameworks, creating exploitable gaps and thereby enabling “jurisdiction-shopping” – where criminals exploit the weakest regulatory link to launder money, evade detection, and undermine the entire system’s integrity.

That begins to change. Europe’s Anti-Money Laundering Regulation (AMLR), set to take effect by 2027, will harmonize AML rules across all EU member states, eliminating the regulatory arbitrage that criminals currently exploit and introduce the Anti-Money Laundering Authority (AMLA), a single European supervisor with the power to coordinate cross-border investigations and sanction noncompliance directly.

At the same time, eIDAS 2.0 will introduce European Digital Identity Wallets, allowing citizens and businesses to verify identity, age and credentials across borders with a single secure credential. This could significantly limit impersonation, account takeover, and cross-border onboarding fraud, which are long-standing weak points in the current system.

IDnow is already putting this into practice. As a participant in EUDI wallet pilot initiatives such as APTITUDE and WE BUILD, two of the EU’s Large-Scale Pilots testing the European Digital Identity Wallet in real-world conditions, IDnow is helping shape what cross-border, interoperable identity actually looks like on the ground. And for the emerging era of attribute-only, privacy-preserving verification, IDnow is ready through its Trust Services offering and provides eIDAS 2-certified Qualified Electronic Attestations of Attributes (QEAAs) that allow organizations to verify specific credentials. “With QEAAs, we can verify anything from a residential address to a professional qualification or an authorization status without collecting more personal data than necessary. And with selective disclosure, users stay in control of exactly what they share with whom, which takes this a step further. Together, that gives our customers something rare: the ability to build genuine trust with their end users while actually reducing friction,” says Uwe Pfizenmaier, Director Product Management at IDnow.

3. Technology that matches the threat

For 2025 and 2026, critical technological defenses have emerged:

Advanced biometric and liveness detection that defeats deepfake attacks using multi-modal verification (like face movement, depth, and sound) and micro-expression analysis, 3D depth mapping or passive liveness checks. IDnow currently offers both passive and active liveness, and its AI-powered solutions detect artifacts of deepfakes like fixed pixels or anomalies in a video stream.
AI-powered document verification that analyzes security markers, like optical variable devices (OVDs), to detect sophisticated fake IDs generated by AI tools and dark web services. IDnow’s Protect capability combats these threats by collecting and connecting signals across device, network and identity data to detect fraud patterns, such as reused document templates, repeated faces, or suspicious session volumes. Additionally, IDnow’s 360 Signals capability detects repeat offenders by identifying suspicious reuse of document templates, amongst other factors.
Real-time behavioural analytics that flag anomalies the moment they appear. Risk Intelligence that is part of IDnow’s Trust Platform analyses behavioural, contextual and technical signals simultaneously, detecting unusual device behaviour, geolocation inconsistencies, VPN masking, mismatched device-IP patterns and signs of emulation or spoofing – all without adding friction for genuine users. Rather than relying on a single check, the system evaluates numerous signals in parallel, building a holistic picture of risk at onboarding, during login or even mid-session.
Continuous monitoring across the full customer lifecycle that detects behavioural changes, suspicious account activity and emerging fraud typologies, including account takeovers and synthetic identities – long after onboarding is complete. IDnow’s Trust Platform centralises signals and alerts in real time, giving institutions the intelligence to act immediately, before fraud escalates.
Synthetic identity detection that cross-references data across multiple sources and timeframes to identify fabricated identities – those created by blending real and fake information to bypass traditional checks. IDnow’s 360 Signals combats this by connecting the dots across sessions: the same biometric template, reused identity information, duplicate documents or recurring device fingerprints, caught in real time, even when each individual signal appears legitimate on its own.
Intelligence and data sharing that leverages anonymized threat data across industries to identify emerging fraud patterns, known fraudster networks and compromised credentials in near real-time. IDnow is currently working on the Fraud Consortium information system, which will allow participating customers to securely share selected fraud-related data, including suspicious IP addresses, email addresses, and document numbers, to strengthen fraud prevention across the network.

The alternative is unacceptable

When criminals industrialize, compliance must innovate. The same AI driving fraud can – and must – power defense. The future lies in AI-guided systems that evolve as fast as threats do, turning reactive defense into proactive prevention.

The tools exist. The frameworks are coming. What’s missing is coordination. Financial institutions must share intelligence even with competitors, while regulators and law enforcement align on faster, privacy-conscious access to data.

“Our ultimate goal is to deliver a flexible, human-assisted fraud prevention solution that puts intelligent automation at its core,” says Kalinowski. “While human oversight remains essential in certain scenarios, the shift toward AI-supported decision-making is clear. Whether our clients require a fully automated fraud engine or a hybrid, human-in-the-loop approach – we adapt to specific business needs, ensuring both security and a seamless user experience.”

The question isn’t whether we have the tools to fight back, but whether we’ll use them. Together.

Fraud doesn’t operate in silos. Your defenses shouldn’t either.

Explore the Trust Platform

Author