Incode Technologies, a global leader in identity security and fraud prevention, has introduced Deepsight, an AI-powered defense system designed to accurately detect and block deepfakes, injected virtual cameras, and synthetic identity attacks.
Deepsight employs multi-modal AI to analyze video, motion, and depth data, identifying inconsistencies that cannot be replicated by synthetic media. This process is completed in under 100 milliseconds without adding friction to user experiences.
Ricardo Amper, Founder and CEO of Incode, stated that deepfakes have evolved into a significant fraud weapon, emphasizing that the ability to fake identity can compromise trust. He added that Deepsight aims to restore this trust by verifying that captures originate from human users in front of a camera, not deepfakes. Deepsight is part of Incode’s broader investment in frontier AI research for identity and trust, which includes Agentic Identity, connecting verified humans with AI agents acting on their behalf.
Deepsight’s assessment of identity is structured across three layers to provide detailed visibility into threats, including identifying the generative model used for fake content. The Behavioral Layer detects subtle interaction anomalies from AI bots or fraud farms. The Integrity Layer verifies camera and device authenticity to block virtual media. The Perception Layer distinguishes deepfakes from genuine human users through AI analysis across multiple capture modalities, such as video, motion, and depth.
Roman Karachinsky, Chief Product Officer at Incode, commented on the growing challenge of verifying human authenticity and confirmed Deepsight’s effectiveness in both laboratory and real-world environments.
The system’s models were benchmarked in Purdue University’s October 2025 study, “Fit for Purpose? Deepfake Detection in the Real World.” This study evaluated 24 detection systems from commercial, government, and academic providers. Incode achieved the highest accuracy and the lowest false acceptance rate among commercial tools, surpassing both government and academic models. Shu Hu, assistant professor at the School of Applied and Creative Computing and Director of the Purdue Machine Learning and Media Forensics (M2) Lab, stated that Incode’s detector demonstrated stronger robustness and reliability in challenging real-world scenarios. Internally, Deepsight was found to be 10 times more accurate than trained human reviewers.
Voi, a European micromobility company, utilizes Incode for AI-powered identity verification and deepfake detection within its fraud solutions. Chris Hobbs, Senior Category Manager, Indirect Procurement at Voi, noted that Incode assists in preventing fraud and ensuring customers meet legal age and safety requirements, particularly given the ease with which deepfakes can be created today.
Deepsight is currently available through the Incode Identity Platform, protecting enterprises across various applications including KYC onboarding, step-up verification, authentication, workforce access, and age verification. It is deployed at organizations such as TikTok, Scotiabank, and Nubank, having protected millions of users across more than six million live identity sessions.
Incode Technologies specializes in trust and identity, aiming to power a world of trust at the speed of AI. The company supports eight of the top ten U.
S. banks, seven of the top eight telecom providers, and numerous leading fintechs, marketplaces, and governments globally, processing over four billion identity checks annually.
