Honor unveiled two new artificial intelligence (AI) innovations at the Mobile World Congress (MWC) Shanghai 2024, on Monday. The smartphone maker introduced an AI Defocus Eye Protection technology that can reduce the risk of myopia or nearsightedness that occurs from staring at a display for too long. It also showcased an AI-powered Deepfake Detection technology that can detect videos that are either made or altered using synthetic means and AI. Interestingly, the company said that both features will be implemented on-device instead of being powered via cloud servers.
Honor unveils new AI technologies
According to a newsroom post by MWC Shanghai, the Chinese consumer technology brand showcased new on-device AI innovations aimed at creating user safety. The announcements were made during Honor’s keynote session at the event titled “The Human-AI Synergy: Intelligent Devices Will Empower People Better”. Alongside introducing the new technological advancements, the company also highlighted the need to build human-centric on-device AI features which are accelerated by dedicated hardware to protect user privacy.
Honor’s AI Defocus Eye Protection
Photo Credit: MWC Shanghai/Honor
The AI Defocus Eye Protection is designed to tackle the global rise in cases of nearsightedness which is caused by long-duration screen usage, the company said. Various studies have shown the effectiveness of peripheral defocus lenses in controlling myopia among individuals. These glasses create an altered visual perception that slows down the eye elongation process responsible for the eye disorder. Now, Honor’s technology uses AI to simulate defocus glasses to a smart device’s display.
As per the company, the AI Defocus Eye Protection can decrease users’ transient myopia by 13 degrees on average after using it for 25 minutes. Citing research, the brand said some users have even experienced a maximum reduction of 75 degrees.
Honor’s AI Deepfake Detection
Photo Credit: MWC Shanghai/Honor
Deepfakes have also been a major problem in recent years, accelerated by the introduction of video-generation AI models. Many AI scientists, tech firms, and regulatory bodies have acknowledged its role in spreading misinformation and manipulating people. Honor’s AI Deepfake Detection is an on-device technology that can analyse videos frame-by-frame to check for information such as “eye contact, lighting, image clarity, and video playback that are imperceptible to the human eyes.”
The AI technology was trained using a large dataset of videos and images relating to online scams. The company claims that it can perform identification, screening, and comparison within three seconds. The feature can display a risk warning if a deepfake is detected. Honor says this can help protect people from falling for scams.
