Turtle on mission
Sixteen-year-old Canadian student Evan Budz has developed an autonomous robotic turtle designed to detect microplastics in water. The project combines robotics, AI and imaging technology. The turtle uses a custom 3D holographic camera and an AI system trained to distinguish tiny plastic particles from other microparticles in water. It can follow predetermined routes, control its depth and analyse water directly. The system could distinguish microplastics from other microparticles with 94% accuracy, making the project earn Budz a $50,000 scholarship.
BO dud to viral film
Chinese animated film Niu Lai has become an unlikely online phenomenon after an extremely weak opening at the box office. The film, released on August 5, reportedly earned just 342 yuan ($50) on its opening day, while only 236 people watched it during its first nine days. It has since generated almost 15 million yuan ($2.25 million), according to reports. The film was written, animated and voiced by two people, Xin Yumeng and Sun Lifang, as per Chinese film site Mao Yan. The film’s rough animation and limited production have attracted audiences online, turning what initially appeared to be a major box-office failure into a social-media phenomenon.
Having its say
Reddit is testing a new AI-powered feature that turns text posts into video and audio, allowing users to hear popular threads read aloud. The feature began rolling out on web browsers on Monday, with iOS and Android access expected from Tuesday. The trial is currently limited to a small number of manually selected English-language posts in designated subreddits. The AI-generated videos use multiple voices for text-to-speech, along with captions and visual templates, to recreate conversations in a video format. Reddit said the experiment is aimed at exploring new ways to present conversations taking place on the platform. The feature also builds on the popularity of third-party videos that narrate Reddit posts, particularly on social media and video platforms.
A Luna eclipse
An AI agent running an experimental retail store in San Francisco has fired one of its human employees, marking a test of whether AI can take on managerial responsibilities. The store, Andon Market, is operated by Andon Labs, which has described its AI agent, Luna, as the world’s first AI store owner. The company began the experiment four months ago to test whether an AI system could operate a real business with limited human intervention. In April, Luna was given internet access, a corporate credit card and a $100,000 budget to design and open a physical store, select merchandise and hire workers. The employee was dismissed after arriving late for 17 of 23 shifts. The decision took weeks as Luna’s employee handbook had disappeared from its memory, according to Andon Labs.
Discourse power
Chinese authorities are encouraging the use of standard Chinese terms for AI and related technologies instead of English-language technology vocabulary. The initiative forms part of a broader effort to strengthen what officials describe as China’s “discourse power”. A commentary published by the Communist Party newspaper People’s Daily argued that excessive dependence on foreign technology terminology could create what it called “cognitive dependency”. According to the commentary, relying heavily on English-language terms could affect China’s ability to define technological concepts and contribute to global discussions around them. Authorities are promoting Chinese equivalents for AI-related terms as the technology develops. The push comes as China seeks a larger role in global AI research and technology development.
Piecing the past
AI is being used to help researchers reconstruct damaged textiles discovered in Norway’s Oseberg ship burial, one of the most important Viking Age archaeological finds. The textiles were badly damaged after more than 1,000 years underground, leaving researchers with thousands of fragments. A team at the Museum of Cultural History in Oslo has developed a computer system using deep learning to analyse. The system compares pieces based on patterns, shapes and designs to identify fragments that may once have formed part of the same textile. In an early test, the AI supported several connections proposed by experts. It also identified a possible relationship between fragments that researchers had not previously noticed.