While browsing the Internet, I came across this Sumsub report on AI-powered fraud. You’re asking, AI-powered fraud? Well, it’s already happening with fake news reports, fake websites that take your identity and bank account details, and more. And with more than a billion people using AI, it’s no wonder that AI-powered fraud has reached industrial scale in 2025.
The days of crude digital fraud, like poorly edited IDs and basic forgeries, are over. A new, professional AI-driven fraud ecosystem has taken hold. Experts call this the ‘Sophistication Shift.’ In 2025, 4.18% of all digital identity verification attempts were flagged as fraudulent. That’s one in every 25 attempts.
The financial impact of AI-powered fraud is massive.
Global identity fraud costs are set to top $50 billion in 2025. In the U.S alone, AI-enabled fraud losses are expected to jump from $12.3 billion in 2023 to $40 billion by 2027.
The Sumsub Identity Fraud Report 2025-2026, based on over 4 million fraud attempts, shows a shift from high-volume, low-skill attacks to fewer but more damaging incidents. Verification systems now block basic schemes, but fraudsters are using accessible AI tools to launch sophisticated, industrial-scale operations.
The Rise of the AI Fraud Ecosystem
In 2025, AI has moved from being a helper tool to becoming an autonomous engine for fraud.
Deepfake Weaponization: New text-to-video systems like OpenAI’s Sora and Google’s Veo can create realistic facial micro-expressions and lighting. Attackers use these tools to stage deepfake liveness checks that closely mimic real human reactions. Visual verification is now a weak point.
Synthetic Identities: Criminals are now creating new digital personas instead of just stealing data. Synthetic identity fraud, which uses AI-generated selfies and fake personal data, jumped 142% year over year in Asia-Pacific.
Industrialized Forgery: Fraud-as-a-service platforms now sell ‘production kits’ that let anyone create thousands of high-quality fake IDs. These forgeries include accurate fonts, holograms, and textures.
Global Landscape: A Tale of Two Realities
The report notes that while overall fraud rates are steady or falling in mature markets, the complexity of attacks is rising fast.
In the United States, the overall fraud rate fell to 1.4%—a 19% decrease from the previous year—thanks to widespread liveness-driven verification. However, deepfake attempts in the U.S. surged by a staggering 237%. Similarly, Canada saw a 10% decline in volume but a 124% increase in deepfake incidents.
Emerging markets are seeing a different trend. In Africa, deepfake-related attempts jumped 367% in the Democratic Republic of the Congo and over 300% in Malawi and Tanzania. In Latin America, Brazil now accounts for nearly 39% of all deepfakes detected in the region.
2026 Forecast: The Arrival of AI Fraud Agents
Looking ahead to 2026, the report warns about AI Fraud Agents. These autonomous systems use generative content and behavioral mimicry to complete verification attempts from start to finish, with no human involved.
“We are witnessing a fundamental shift,” says Andrew Sever, CEO at Sumsub. “The next frontier of fraud prevention will belong to those who can unite human insight, data intelligence, and AI precision to build trust at scale”.
The New Defense: Behavioral Intelligence
With forged documents and videos now common, the industry is turning to behavioral AI. Deepfakes may look real, but digital traces like mouse movements, typing patterns, and app navigation often reveal automation. Multi-layer behavioral models are now key to spotting inconsistencies that static ID checks miss.
Fraud is now more deliberate and multi-layered. For businesses, staying audit-ready and compliant means adapting to an AI-powered arms race that is only accelerating. As companies become more proactive in addressing AI-powered fraud, consumers can likely expect new methods and approaches to protect their digital identities and financial information.
Read the whole Sumsub Identity Fraud Report 2025-2026 here.