Modern technology allows today’s digital customers to use instant sign up procedures and applications for accurate, safe, and fast remote identity verification. However, it can be difficult to verify whether a user is genuine or not.
DuckDuckGoose’s deepfake detection software protects your company from fraudsters seeking to deepfake someone elses identity and facial biometrics to pose as their own. By using cutting-edge deepfake detection, we ensure that the right people are signing up for your application or service.
Fraudsters managed to use facial images bought on the black market to generate a synthetic identity. They used this to create a business that provided forgered tax invoices worth $76.2 million dollars (USD).
A simple application was used to manipulate the pictures to create a deepfake video in which the face could blink, nod, and open its mouth. Then, a mobile camera was used to carry out the facial recognition process, allowing them to deceive the tax invoice system into trusting the deepfake.
A recent study conducted by researchers from South Korea looked into APIs from Microsoft and Amazon who claim to provide highly precise commercial facial recognition web services for various softwares in order to meet the end-user requirements.
Their goal was to illustrate the vulnerability of facial recognition systems by fooling the algorithm into falsely recognising images as a celebrity. What they found was that 78% of the targeted celebrity the researchers put into Microsoft’s Azure Cognitive Service was successfully deceived, while Amazon deceived 68.7%.