Honor unveiled two new synthetic intelligence (AI) improvements on the Cell World Congress (MWC) Shanghai 2024, on Monday. The smartphone maker launched an AI Defocus Eye Safety know-how that may scale back the danger of myopia or nearsightedness that happens from watching a show for too lengthy. It additionally showcased an AI-powered Deepfake Detection know-how that may detect movies which might be both made or altered utilizing artificial means and AI. Apparently, the corporate mentioned that each options will probably be carried out on-device as an alternative of being powered through cloud servers.
Honor unveils new AI applied sciences
Based on a newsroom submit by MWC Shanghai, the Chinese language shopper know-how model showcased new on-device AI improvements geared toward creating consumer security. The bulletins had been made throughout Honor’s keynote session on the occasion titled “The Human-AI Synergy: Clever Gadgets Will Empower Folks Higher”. Alongside introducing the brand new technological developments, the corporate additionally highlighted the necessity to construct human-centric on-device AI options that are accelerated by devoted {hardware} to guard consumer privateness.
Honor’s AI Defocus Eye Safety
Picture Credit score: MWC Shanghai/Honor
The AI Defocus Eye Safety is designed to sort out the worldwide rise in circumstances of nearsightedness which is brought on by long-duration display utilization, the corporate mentioned. Numerous research have proven the effectiveness of peripheral defocus lenses in controlling myopia amongst people. These glasses create an altered visible notion that slows down the attention elongation course of answerable for the attention dysfunction. Now, Honor’s know-how makes use of AI to simulate defocus glasses to a sensible system’s show.
As per the corporate, the AI Defocus Eye Safety can lower customers’ transient myopia by 13 levels on common after utilizing it for 25 minutes. Citing analysis, the model mentioned some customers have even skilled a most discount of 75 levels.
Honor’s AI Deepfake Detection
Picture Credit score: MWC Shanghai/Honor
Deepfakes have additionally been a serious downside in recent times, accelerated by the introduction of video-generation AI fashions. Many AI scientists, tech companies, and regulatory our bodies have acknowledged its position in spreading misinformation and manipulating individuals. Honor’s AI Deepfake Detection is an on-device know-how that may analyse movies frame-by-frame to verify for data resembling “eye contact, lighting, picture readability, and video playback which might be imperceptible to the human eyes.”
The AI know-how was educated utilizing a big dataset of movies and pictures regarding on-line scams. The corporate claims that it may possibly carry out identification, screening, and comparability inside three seconds. The characteristic can show a danger warning if a deepfake is detected. Honor says this can assist shield individuals from falling for scams.