分类: technology

  • Regulator contacts Meta over workers watching intimate AI glasses videos

    Regulator contacts Meta over workers watching intimate AI glasses videos

    The UK Information Commissioner’s Office (ICO) has initiated formal communications with Meta following disturbing revelations about the handling of sensitive user content captured through the company’s AI-powered smart glasses. This development comes after a joint investigation by Swedish publications Svenska Dagbladet and Goteborgs-Posten uncovered that outsourced workers in Kenya had access to highly personal recordings, including intimate moments and private activities.

    According to the investigation, data annotators employed by Nairobi-based outsourcing firm Sama reviewed unfiltered content from Meta’s Ray-Ban smart glasses, which included footage of individuals using toilets, engaging in sexual activity, and changing clothes. One worker reportedly stated, ‘We see everything – from living rooms to naked bodies,’ highlighting the severe privacy implications.

    Meta acknowledged that contracted workers sometimes review user content to enhance the AI experience but emphasized its commitment to data protection. ‘When people share content with Meta AI, like other companies we sometimes use contractors to review this data to improve people’s experience with the glasses,’ the company told BBC News. Meta claims to employ privacy filters, including face blurring technology, though sources indicated these measures frequently failed.

    The Ray-Ban Meta glasses feature recording capabilities activated either manually or through voice commands, with a visible light indicator signaling active recording. However, users may not fully comprehend that their content could undergo human review, as detailed in Meta’s extensive privacy policies that few thoroughly read.

    The ICO expressed significant concerns, stating: ‘Devices processing personal data, including smart glasses, should put users in control and provide for appropriate transparency. Service providers must clearly explain what data is collected and how it is used.’ The watchdog is now seeking clarification on Meta’s compliance with UK data protection laws.

    This incident raises broader questions about the ethical implications of AI-powered wearable technology and the responsibility of tech giants to protect user privacy amidst rapid technological advancement. Previous reports have already documented concerns about smart glasses being used for non-consensual recording, particularly affecting women.

    Sama, the outsourcing company involved, has a complicated history with content moderation services. Previously designated as an ‘ethical’ B-corp, the company faced criticism and legal action from former employees regarding its content moderation practices and has since discontinued such services.

  • China achieves breakthrough in high-orbit satellite-ground laser communication

    China achieves breakthrough in high-orbit satellite-ground laser communication

    Chinese scientists have established a groundbreaking milestone in space communication technology by successfully demonstrating high-speed laser data transmission between Earth and a satellite in high orbit. The unprecedented achievement, accomplished across a staggering distance of 40,740 kilometers, represents a quantum leap in satellite-ground communication capabilities.

    The pioneering experiment, a collaborative effort between the Chinese Academy of Sciences’ Institute of Optics and Electronics, Beijing University of Posts and Telecommunications, and the China Academy of Space Technology, utilized a specialized observatory facility in Yunnan province to establish a stable optical link with a geosynchronous satellite. This technological marvel achieved simultaneous two-way communication at remarkable speeds of 1 gigabit per second – comparable to terrestrial fiber-optic performance but accomplished across interplanetary distances.

    What distinguishes this breakthrough is its exceptional operational performance: the system established a laser connection in merely four seconds and maintained uninterrupted transmission for over three hours. This represents a monumental improvement from previous minute-level durations to sustained hour-level operation, marking the first time such stability has been achieved in the challenging high-orbit environment.

    The technological implications extend far beyond raw data transmission speeds. This advancement effectively transforms high-orbit satellites from simple data relay stations into potential intelligent processing hubs capable of receiving complex commands and executing sophisticated operations in real-time. The demonstrated capability addresses two critical challenges in space communication: achieving peak transmission rates for data-intensive applications and ensuring prolonged operational stability for advanced interactive systems.

    Researchers emphasize that this breakthrough establishes a mature engineering model for future space infrastructure. The validated technology paves the way for establishing high-speed laser communication networks with lunar bases, Mars missions, and deep space probes, effectively laying the foundation for an integrated Earth-space network that could revolutionize humanity’s approach to space exploration and satellite operations.

  • Some West African farmers turn to TikTok as part of agriculture’s changing image

    Some West African farmers turn to TikTok as part of agriculture’s changing image

    DAKAR, Senegal — A digital agricultural revolution is unfolding across West Africa as farmers leverage social media platforms like TikTok to transform traditional farming practices and market access. What began as entertainment platforms have evolved into vital business tools for millions of agricultural producers throughout the region.

    Pape Fall, a Senegalese farmer, exemplifies this transformation. Initially using TikTok for football highlights and humorous content, Fall now conducts most of his business through the platform. His profile features looped videos showcasing cucumber harvests set to Senegalese rap music, complete with availability details and contact information. This digital approach has enabled him to directly connect with buyers, bypassing traditional intermediaries.

    The phenomenon represents a significant shift in how agriculture is perceived and practiced in a region historically plagued by food insecurity and poverty. While foreign funding reductions from U.S. and other international donors have exacerbated challenges, social media has emerged as an unexpected solution, providing both educational resources and market opportunities.

    According to Abbie Phatty-Jobe, digital agriculture specialist at U.K.-based research firm Caribou, “Social media is democratizing access to information for farmers.” Caribou’s research, the first to examine regional farmers’ TikTok usage, has helped establish a network of 24 agricultural influencers across 11 African countries, reaching approximately 5 million people with content that translates scientific research into practical farming advice.

    The platform’s impact extends beyond marketing. Fall credits TikTok tutorials with improving his cultivation techniques, specifically mentioning advice from a Moroccan farmer about cucumber stem management that significantly increased his yields. This knowledge exchange occurs across continents, with farmers from North Africa to Asia sharing best practices.

    Regional preferences vary significantly: West African farmers favor TikTok’s video format and local language support, while East African producers prefer Facebook’s text-based posts due to higher literacy rates. Successful entrepreneurs like 29-year-old Nogaye Sene have built substantial businesses through these platforms, with Instagram and TikTok generating 70% of her clients for farmland management services targeting both local residents and diaspora investors.

    However, challenges persist. Sene warns of widespread online scams involving fake farming consultants, while researcher Nicolas Paget notes that approximately 80% of farmers lack smartphones capable of running these applications. Even those with access face prohibitive data costs, with West African data packages being more expensive than European equivalents.

    Despite these barriers, the organic adoption of existing social platforms appears more effective than customized solutions. As Paget observes, “Farmers are using existing apps in very creative ways and adapting them to their needs”—a testament to the innovative spirit driving agricultural digitalization in West Africa.

  • X suspends revenue sharing for undisclosed AI war videos

    X suspends revenue sharing for undisclosed AI war videos

    In a significant policy shift, social media platform X has announced stringent measures against creators who disseminate artificially generated conflict footage without proper disclosure. The Elon Musk-owned company revealed Tuesday that participants in its Creator Revenue Sharing program will face 90-day suspensions for posting AI-generated videos of armed conflicts without clear labeling.

    The decision emerges amid escalating geopolitical tensions involving the United States, Israel, and Iran, where synthetic media poses unprecedented challenges to information integrity. Nikita Bier, X’s Head of Product, emphasized the critical need for authentic battlefield information during wartime, noting that current AI technologies have made it “trivial to create content that can mislead people.”

    This policy reversal marks a notable departure from X’s previous stance on content moderation. Since Musk’s $44 billion acquisition of the platform (formerly Twitter) in October 2022, the company has systematically dismantled most misinformation policies, characterizing them as forms of censorship. The new framework introduces escalating penalties, with repeat offenders facing permanent removal from the revenue program that distributes advertising earnings to eligible creators.

    Enforcement will leverage multiple detection methods, including Community Notes (the platform’s crowd-sourced fact-checking system), metadata analysis, and technical signals embedded within AI-generated content. The company confirmed ongoing refinements to both policies and technical infrastructure to enhance trust during critical global events.

  • Goodbye officer, hello touchscreen – a vision for policing in Ethiopia

    Goodbye officer, hello touchscreen – a vision for policing in Ethiopia

    Ethiopia has inaugurated its first unmanned police station in Addis Ababa’s Bole district, marking a significant milestone in the country’s Digital Ethiopia 2030 initiative. The innovative facility features partitioned booths equipped with computer tablets instead of traditional front desks, creating an unusually quiet environment devoid of the typical commotion of police stations.

    Commander Demissie Yilma, head of the police technology expansion department, demonstrated the system to BBC reporters. Users select incident categories—crime, traffic reports, or general concerns—enter details via tablet, and subsequently connect with remote officers via video call rather than chatbots. This hybrid approach combines digital automation with human oversight, ensuring immediate response coordination with patrol units when necessary.

    The pilot station recorded modest initial usage with just three reports in its first week—a lost passport, financial fraud case, and routine complaint. Authorities anticipate increased adoption as public awareness grows. Prime Minister Abiy Ahmed has championed the initiative as part of broader efforts to make ‘law enforcement institutions competent and competitive’ through digital reform.

    This innovation addresses both accessibility challenges and personnel constraints, potentially expanding police services to underserved areas. However, it emerges against a complex backdrop: Ethiopia trails continental peers in digital transformation due to historically low internet penetration rates and periodic connectivity disruptions stemming from recent conflicts and political instability.

    The national digital strategy aims to overhaul public services including identity systems, payments, and judicial administration. Recent telecommunications sector liberalization has accelerated mobile payment adoption in the local currency, while government services migration online continues progressively.

    Technology policy analyst Zelalem Gizachew notes measurable progress: ‘Digital payments have boomed with trillions of birr moving through electronic transactions. Broadband access has expanded sharply, and over 130 government services have been digitized.’

    Yet significant challenges persist. UNESCO data indicates 79% of Ethiopians lack internet access, creating adoption barriers. Senior software engineer Birhan Nega Cheru acknowledges efficiency gains—reduced paperwork and office visits—but highlights cybersecurity vulnerabilities and exclusion risks for digitally illiterate, elderly, rural, and low-income populations.

    The station remains an experimental prototype operating in a controlled environment with officer assistance. Its future expansion depends less on technological sophistication than organic public adoption without guided facilitation. This quiet room in Bole represents both a practical test case and symbolic manifestation of Ethiopia’s broader digital ambitions intersecting with everyday civic life.

  • Meet ‘Dasheng’: A system-level scientific AI agent

    Meet ‘Dasheng’: A system-level scientific AI agent

    Shanghai, March 1, 2026 – In a landmark development for artificial intelligence research, the Shanghai Academy of AI for Science and Fudan University have jointly launched a significantly upgraded version of their NovaInspire platform, featuring the groundbreaking scientific AI agent ‘Dasheng’. Named after the legendary Monkey King from Chinese mythology, this system represents a quantum leap in AI capabilities for scientific discovery.

    The Dasheng AI agent integrates multiple cutting-edge technologies including multimodal foundation models, long-term collective memory systems, and expert-level scientific capabilities within a secure, trustworthy framework. What distinguishes Dasheng from conventional AI systems is its capacity for multi-threaded processing and continuous autonomous evolution, enabling it to tackle complex, real-world scientific problems that have previously challenged researchers.

    This system-level scientific partner incorporates self-driving laboratory capabilities, allowing for unprecedented automation in experimental processes and data analysis. The integration of these advanced features positions Dasheng as a transformative force in research methodology, potentially accelerating breakthroughs across multiple scientific disciplines from materials science to pharmaceutical development.

    The development team emphasizes that Dasheng represents more than just incremental improvement – it constitutes a fundamental paradigm shift in how scientific research can be conducted. By combining artificial intelligence with laboratory automation, the system creates a continuous cycle of hypothesis generation, experimentation, and analysis that operates with minimal human intervention.

    This announcement comes at a time when global competition in artificial intelligence research has intensified, with multiple nations and corporations investing heavily in next-generation AI systems. The successful development of Dasheng demonstrates China’s growing capabilities in advanced AI research and its application to scientific challenges.

  • China issues guidelines to boost recycling, utilization of solar modules

    China issues guidelines to boost recycling, utilization of solar modules

    In a significant move toward sustainable energy management, China has launched comprehensive national guidelines to establish a robust ecosystem for recycling and reusing decommissioned solar photovoltaic (PV) modules. The Ministry of Industry and Information Technology, in coordination with five other governmental bodies, introduced the framework on Tuesday to address the impending surge of retired solar equipment expected from the nation’s massive renewable energy expansion.

    The newly released directives outline a multi-faceted approach to building system-level capacity for handling end-of-life PV components. Central to the strategy is the enhancement of green manufacturing standards for solar modules, increased incorporation of recycled materials in production, and the development of improved scrap assessment protocols alongside advanced testing methodologies.

    Technological innovation forms the cornerstone of China’s recycling initiative. The guidelines emphasize critical breakthroughs in several key areas including surface-structure dismantling techniques, high-efficiency separation of laminated components, and advanced material extraction processes. These technological advancements will enable the broader application of recycled PV-derived products across multiple industrial sectors such as metal smelting, equipment manufacturing, and construction materials.

    The policy framework establishes concrete targets, including achieving cumulative comprehensive utilization of retired PV modules reaching 250,000 tonnes by 2027. By 2030, China aims to realize substantial upgrades in recycling technology and equipment, strengthen industrial innovation capabilities, and expand application scenarios for recycled products. The long-term vision includes creating a tightly coordinated upstream-downstream supply chain with rationally distributed capacity布局 capable of handling large-scale retirements of solar infrastructure.

    Additionally, the guidelines call for the development of comprehensive technical standards for green design and utilization of PV modules while fostering a cohort of leading enterprises specializing in end-of-life solar component processing. This strategic move positions China at the forefront of circular economy practices within the global renewable energy sector.

  • NPC deputy calls for development of ‘AI plus manufacturing’

    NPC deputy calls for development of ‘AI plus manufacturing’

    Zhang Fan, Vice-General Manager of China Electrical Equipment Group and National People’s Congress deputy, has positioned artificial intelligence integration with manufacturing as a national priority requiring comprehensive strategic planning. The prominent industry leader argues that China’s manufacturing sector must embrace AI technologies to achieve transformative innovation and maintain global competitiveness.

    During recent parliamentary sessions, Zhang emphasized that ‘AI plus manufacturing’ represents the future of industrial development, necessitating strengthened top-level design and clear industry guidelines. His advocacy builds upon previous policy recommendations submitted last year, which called for deeper integration between artificial intelligence and manufacturing processes to drive high-quality development.

    Zhang’s proposals have gained significant traction within China’s policy-making circles, receiving substantial attention from relevant government departments. His position combines practical industry experience with legislative influence, as demonstrated during his January 2026 inspection of China Electrical Equipment Group’s research facilities in Shanghai, where he assessed technological readiness for AI implementation.

    The manufacturing expert’s recommendations align with China’s broader technological ambitions, suggesting that systematic AI adoption could revolutionize production efficiency, product quality, and innovation capabilities across traditional manufacturing sectors. This approach mirrors global trends where advanced economies are increasingly leveraging AI to maintain industrial leadership.

    Zhang’s dual role as both industry executive and elected representative provides unique insights into the practical challenges and opportunities presented by AI integration, making his policy recommendations particularly valuable for China’s ongoing technological modernization efforts.

  • India’s top court angry after junior judge cites fake AI-generated orders

    India’s top court angry after junior judge cites fake AI-generated orders

    India’s judicial system faces a critical test of technological integration as the Supreme Court intervenes in a landmark case involving artificial intelligence-generated legal judgments. The controversy emerged when a junior civil judge in Vijaywada, Andhra Pradesh, utilized AI to produce four fabricated legal citations during a property dispute ruling in August last year.

    The Supreme Court has characterized this incident as a matter of ‘institutional concern’ that directly impacts ‘the integrity of adjudicatory process.’ In a stern response, the court stayed the lower court’s ruling and differentiated between human error and technological misconduct, stating that reliance on AI-generated falsehoods constitutes more than mere mistake—it represents professional misconduct.

    The case reached India’s highest judicial authority after defendants successfully demonstrated that the referenced legal precedents were entirely fictitious. Although the Andhra Pradesh High Court had previously acknowledged the AI-generated nature of the citations, it upheld the original decision, accepting the judge’s claim that she acted in ‘good faith’ during her first experience with AI tools.

    This incident reflects a global pattern of judicial systems struggling with AI integration. Recent months have witnessed similar controversies in United States federal courts and England’s High Court, where lawyers presented fabricated AI-generated case materials. The international legal community increasingly recognizes that generative AI systems frequently ‘hallucinate’—creating convincing but entirely fictional legal references and sources.

    India’s Supreme Court has responded by issuing notices to the country’s Attorney General, Solicitor General, and the Bar Council of India, signaling potential regulatory developments. This action follows the court’s recent publication of a white paper on AI in judiciary, which emphasized the necessity of maintaining human oversight and institutional safeguards despite technological adoption.

    The case represents a pivotal moment in legal technology regulation, balancing efficiency gains against fundamental requirements for judicial accuracy and integrity.

  • Radar system aids rail tunnel safety study

    Radar system aids rail tunnel safety study

    A groundbreaking radar technology developed by researchers at Xidian University is transforming railway tunnel safety inspections across China. The innovative system, mounted on moving trains, recently completed comprehensive scans of 77 railway tunnels in just eight hours—setting an unprecedented industry speed record.

    Professor Su Tao’s team at the university’s Hangzhou Institute of Technology engineered the radar equipment to operate at speeds up to 160 kilometers per hour, representing a 30-fold increase in efficiency compared to conventional inspection methodologies. This technological leap addresses critical maintenance challenges within China’s rapidly expanding rail network, which encompasses 17,177 operational railway tunnels spanning 22,669 kilometers as of late 2023.

    The radar system specifically targets hidden structural defects that develop over time, including internal voids, lining delamination, and water infiltration—flaws that traditionally required manual inspection during brief intervals between train operations. Project leader Xu Zhi emphasized the limitations of conventional approaches: “Traditional methods are slow, risky, and produce few traceable digital records. For a network measuring tens of thousands of kilometers, routine inspection becomes impractical.”

    Through four years of on-site development in active tunnels, the research team overcame significant technical hurdles, including adapting radar signals for arched tunnel environments, developing pattern recognition algorithms for defect identification, and filtering noise generated by train vibrations. The system performs full cross-section scans from distances exceeding 4.5 meters, eliminating the need for physical contact with tunnel walls.

    Validation procedures involving core sampling at radar-identified locations demonstrated remarkable accuracy, with physical evidence consistently confirming the radar’s diagnoses of thin lining and internal voids. The technology now progresses toward automated data processing and three-dimensional visual reconstruction capabilities.

    “High precision rapid inspection is only the first step,” Xu noted. “We’re developing systems that will enable inspectors to view tunnel interiors on screens with defect locations, types, and scales clearly visible—essentially providing complete CT scans of tunnel infrastructure during normal train operations.”