Hiya, a Seattle-based technology company, has introduced a new AI-powered assistant within its Hiya AI Phone app designed to combat deepfake voice scams in real-time. The application provides comprehensive call screening capabilities and synthetic voice detection to protect users from fraudulent calls, addressing a growing threat that security experts have warned about since 2020.
The app continuously analyzes call audio using advanced AI models trained to detect scam language and other suspicious indicators during conversations. When potential scams are identified – whether from a human caller, recording, or AI-generated voice – users receive immediate alerts. The development arrives as security agencies worldwide have reported increasingly sophisticated AI-powered fraud schemes.
Recent survey data from Hiya and Censuswide reveals the scope of the problem: one in three Americans reported receiving at least one deepfake scam call, with 45 percent of targets falling victim and losing an average of $7,200. These findings match broader industry research showing that consumers often overestimate their ability to detect deepfake content.
The application emphasizes user privacy through encrypted call data storage on the device, with data being deleted seconds after download. While available as a free download for both Android and iPhone users, premium features are offered via a $9.99 monthly subscription with a two-week trial period. The secure data handling approach reflects growing industry standards for carrier-level authentication and security.
Hiya has enhanced its AI Voice Detection capability with specific defenses for handling background sounds, using proprietary toolkits to generate training data that reflects real-world scenarios. The system employs deep learning models trained on thousands of hours of audio, achieving over 99 percent accuracy in both controlled and real-world environments. Detection begins within one second of speech, building on advances in voice biometric authentication technology.
The technology is designed to provide protection regardless of the synthesis tool or language used, offering multi-channel compatibility. The solution can be deployed via cloud-based APIs or on-premises systems to accommodate various carrier requirements, similar to approaches being adopted by other major telecommunications fraud prevention initiatives.
Sources: BestofAI, Hiya, Help Net Security, Hiya Blog
Follow Us