HONOR has announced two groundbreaking on-device AI technologies designed to enhance user empowerment and safety. Unveiled during its keynote at MWC Shanghai 2024, titled “The Human-AI Synergy: Intelligent Devices Will Empower People Better,” HONOR introduced its latest human-centric AI advancements: AI Defocus Eye Protection and AI Deepfake Detection. The keynote also emphasized HONOR’s commitment to human-centric on-device AI and the potential of AI-enabled hardware to deliver personalized services while safeguarding user privacy.
“AI is revolutionizing our lives and driving the smartphone industry forward, but much of the industry has focused on cloud-based AI, which is just a part of the puzzle. On-device AI, which is run on smartphones that understand us better than any other devices, is uniquely positioned to deliver services that are tailored to us and our preferences,” said George Zhao, CEO of HONOR Device Co., Ltd. “At HONOR, we believe on-device AI has the potential to empower users and become more capable in everything they do. We invite all industry players to join us as we explore the untapped potential of on-device AI.”
Evolving Eye Protection Technology from Prevention to Relief
HONOR introduced AI Defocus Eye Protection at MWC Shanghai, addressing the global rise in nearsightedness caused by prolonged screen time. This AI technology simulates defocus glasses on smart device displays to promote eye health. Defocus glasses work by inducing controlled defocus in the peripheral visual field, maintaining clear central vision and slowing down eye elongation, a key factor in nearsightedness.
Moving beyond prevention, HONOR’s AI Defocus Eye Protection provides relief by decreasing users’ transient myopia by an average of 13 degrees after 25 minutes of reading, with some users experiencing up to a 75-degree reduction.
Mitigating Fraud with AI Deepfake Detection
HONOR has developed AI Deepfake Detection to combat the risks posed by cloud-based Deepfake AI technology. This feature analyzes frame-by-frame details such as eye contact, lighting, image clarity, and video playback to identify subtle flaws undetectable by the human eye. Trained on an extensive dataset of videos and images related to online scams, the AI can identify, screen, and compare content within three seconds. Upon detecting synthetic or altered content, it issues a risk warning, alerting users and preventing further interaction with potential scams for their safety.
