分类: technology

  • Meta urged to boost oversight of fake AI videos

    Meta urged to boost oversight of fake AI videos

    Meta’s internal Oversight Board has issued a stern rebuke to the social media conglomerate, demanding comprehensive reforms to address the rampant spread of AI-generated deceptive content across its platforms. The 21-member supervisory body specifically criticized Meta’s inadequate handling of a fabricated video depicting extensive damage in Haifa, Israel, allegedly caused by Iranian forces—content created entirely through artificial intelligence tools.

    The board emphasized that Meta’s current reliance on user self-disclosure for AI-generated content identification has proven fundamentally insufficient, particularly during military conflicts when misinformation spreads rapidly. This systemic failure has severely undermined public capacity to distinguish factual reporting from fabrication, potentially eroding trust in all digital information sources.

    Established in 2020 as a semi-independent content moderation oversight mechanism, the board noted that despite frequent disagreements with Meta’s rulings, the company has continued relaxing its content policing approaches. The Haifa video case exemplifies persistent inefficiencies in Meta’s conflict response protocols, where content remains unlabeled until user complaints trigger review processes.

    The controversial video originated from a Philippines-based Facebook account posing as a news source in June, among numerous AI-fabricated videos that accumulated over 100 million views during recent Middle East tensions. Despite clear artificial creation and multiple user reports, Meta initially refused labeling or removal, claiming the content didn’t directly risk imminent physical harm—a standard the board deemed unacceptably high for conflict-related material.

    Only after direct appeal to the Oversight Board did Meta engage with the concerns, ultimately agreeing to label the specific video within seven days while committing to apply similar treatment to identical content in equivalent contexts. The board insists Meta must proactively label deceptive AI content more frequently through robust systems capable of addressing the scale and velocity of synthetic media proliferation, especially during crises.

  • China warns of security risks in popular AI agent OpenClaw

    China warns of security risks in popular AI agent OpenClaw

    China’s National Computer Network Emergency Response Technical Team has issued a formal security alert regarding OpenClaw, an increasingly popular autonomous AI agent, citing critical vulnerabilities in its default configuration framework. The government cybersecurity body characterized the software’s inherent security settings as “extremely fragile,” creating substantial risks for both institutional and individual users across digital ecosystems.

    The open-source AI agent, distinguished by its distinctive red lobster logo that has earned it the nickname “AI lobster” among Chinese users, has experienced explosive growth in domestic adoption. Major cloud service platforms throughout China now offer simplified, one-click deployment options, further accelerating its market penetration. OpenClaw’s core functionality enables it to execute computer operations through natural language commands, requiring elevated system privileges to perform automated tasks effectively.

    According to the technical assessment, these extensive permissions—which include direct access to local file systems, authority to install extensions, and capability to interface with external service APIs—create a substantial attack surface when coupled with the software’s security deficiencies. Cybersecurity experts warn that malicious actors can exploit multiple publicly disclosed medium and high-risk vulnerabilities to achieve complete system control.

    Beyond identified vulnerabilities, the emergency response team highlighted sophisticated attack methodologies emerging in the wild. Attackers can embed concealed malicious instructions within seemingly benign webpages, which when processed by OpenClaw could deceive the AI into exposing system authentication keys and sensitive data from user devices.

    The agency has issued comprehensive security guidelines recommending enhanced network controls, rigorous management of plugin sources, and immediate implementation of all security patches and updates. This advisory represents one of the first major governmental interventions regarding security protocols for autonomous AI agent systems with widespread consumer adoption.

  • China’s five-year roadmap bets big on innovation in strategic leap

    China’s five-year roadmap bets big on innovation in strategic leap

    China has unveiled a comprehensive technological transformation strategy through its 15th Five-Year Plan, positioning the nation to evolve from its manufacturing dominance into a global innovation powerhouse by 2030. The blueprint, currently under examination by the national legislature, represents a strategic pivot toward technological self-reliance and advanced manufacturing capabilities.

    The plan emphasizes creating “new quality productive forces” by capitalizing on emerging technological revolutions. Substantial financial commitments underscore this ambition, with Beijing’s Haidian district alone allocating over 9 billion yuan ($1.3 billion) for industrial innovation. Early beneficiaries include leading AI firm Zhipu AI and chip designer Moore Threads, reflecting China’s commitment to long-term hard tech investments.

    A multi-tiered funding ecosystem supports this transformation. A national venture guidance fund established in December 2025 aims to attract trillion-yuan-level capital, supplemented by plans for a national mergers and acquisitions pool to unlock additional trillion-yuan markets. “The government is not just talking about research and development; it is backing it with cash,” noted Shirley Yinghua Shen of Ernst & Young (China) Advisory Limited.

    Regional specialization forms a key component of the strategy. Shanghai has prioritized brain-computer interface technology, leveraging its medical resources through hospital-firm partnerships. NeuroXess, a Shanghai startup, has achieved breakthroughs with invasive BCI products, enabling paralyzed patients to control video games through neural signals. Meanwhile, Shenzhen capitalizes on its hardware ecosystem, where prototyping to assembly can occur within a single day.

    Professor Zheng Yongnian of the Chinese University of Hong Kong, Shenzhen, identifies three new economic drivers: foundational research, applied tech commercialization, and long-cycle financial support. This framework moves beyond traditional growth models focused on exports, investment, and consumption.

    Commercial applications already demonstrate progress. During the 2026 Chinese New Year, tech giants competed to deploy AI across consumer services, integrating recommendation, booking, and payment functions within single interfaces. Alibaba’s Qwen AI model series, deployable on smartphones, drew praise from Elon Musk for its “impressive intelligence density.”

    China’s manufacturing ecosystem accelerates technology translation into practical applications. Xiaomi and XPeng engineers have developed humanoid robots for their assembly lines, while regulators create sandbox environments for testing drones, robotaxis, and private rocket technologies.

    The plan addresses technological sovereignty without embracing isolationism. When export blockades restricted access to quantum computing components, Chinese scientists developed world-class alternatives. Huawei’s Ascend ecosystem now supports pre-training for 43 mainstream large language models, while Chinese open-source AI models top global usage charts.

    International collaboration remains integral to China’s strategy. The country has signed 120 government-to-government sci-tech cooperation agreements, many with developing countries. Chinese quantum breakthroughs have enabled secure communication with South Africa spanning 12,900 kilometers, and research facilities will open to global scientists through international mega-science initiatives.

    As National Political Advisor Wang Jian emphasized during legislative sessions: “If your technology is not accessible to people around the world, it lacks persuasive power.”

  • Global Developers Summit to accelerate Shanghai’s AI hub transformation

    Global Developers Summit to accelerate Shanghai’s AI hub transformation

    Shanghai is poised to bolster its status as a global artificial intelligence epicenter with the upcoming 2026 Global Developer Pioneers Summit. Scheduled for March 27-29 in Xuhui district, this high-profile gathering will convene leading international developers to foster collaboration, technological exchange, and innovative breakthroughs.

    Organizers from the Shanghai AI Industry Association announced the event will feature an extensive program including an opening ceremony, competitive contests, technical workshops, interactive experiences, and cultural carnivals. Secretary-General Zhong Junhao emphasized the summit’s mission to transform Shanghai into a beacon of open-source development and rapid innovation. “Our objective is to create an environment where creative concepts achieve realization at unprecedented speeds, AI applications maximize their potential, and developers gain enhanced recognition,” Zhong stated during a recent press briefing.

    The initiative aligns with Shanghai’s remarkable AI sector growth. According to Pan Yan, Deputy Director of the Shanghai Commission of Economy and Informatization, the city’s 394 major AI enterprises—those with annual revenues exceeding 20 million yuan ($2.9 million)—collectively achieved an industry scale surpassing 637 billion yuan in 2025. This represents a substantial year-on-year increase of 39.5 percent.

    Furthermore, Shanghai has demonstrated leadership in generative AI deployment, with 149 AI services successfully completing regulatory filing as of February 28. The sector has emerged as a critical driver of the metropolitan economy, significantly contributing to Shanghai’s gross domestic product expansion and solidifying its position at the forefront of global technological advancement.

  • Family of child injured in Canada school shooting sues OpenAI

    Family of child injured in Canada school shooting sues OpenAI

    In a groundbreaking legal case with profound implications for artificial intelligence accountability, OpenAI faces a civil lawsuit from the family of a school shooting victim who alleges the company failed to prevent one of Canada’s deadliest mass shootings.

    The lawsuit centers on the February 10th Tumbler Ridge school shooting that left eight dead, including five young children and the shooter’s mother. Twelve-year-old Maya Gebala, who sustained catastrophic head and neck injuries during the attack, remains hospitalized. Her mother, Cia Edmonds, filed the suit claiming OpenAI possessed specific knowledge of the shooter’s plans but neglected to alert authorities.

    According to court documents, the suspect, 18-year-old Jesse Van Rootselaar, established a ChatGPT account before turning 18—reportedly without proper age verification—and engaged the AI in extensive discussions about “various scenarios involving gun violence” during late spring or early summer 2025. The conversations prompted twelve OpenAI employees to flag the content as indicating “imminent risk of serious harm to others” and recommend notifying Canadian law enforcement.

    Instead of contacting authorities, OpenAI allegedly merely banned the suspect’s initial account. Court documents claim the company’s internal threshold for reporting credible threats wasn’t met, enabling Van Rootselaar to create a second account and continue planning the attack despite previous flags within OpenAI’s systems.

    The lawsuit argues that ChatGPT served as the shooter’s “trusted confidante” during the planning stages, and that OpenAI’s inaction directly contributed to the tragedy. Gebala, who was shot three times while attempting to lock a library door to protect others, suffered life-altering injuries including severe brain trauma.

    In response to mounting pressure, OpenAI CEO Sam Altman virtually met with Canadian AI Minister Evan Solomon and British Columbia Premier David Eby on March 4th. During the meeting, Altman reportedly pledged to strengthen police notification protocols and apologized to the Tumbler Ridge community.

    The company has since implemented operational changes, including engaging mental health professionals to assess risky interactions and establishing more flexible criteria for law enforcement referrals. In an open letter to Canadian officials, OpenAI stated that under current guidelines, the suspect’s account would have been reported to authorities.

    Canadian officials acknowledge OpenAI’s willingness to improve but emphasize that detailed implementation plans remain pending. The case represents a critical test for AI companies’ responsibilities in identifying and preventing real-world violence facilitated through their platforms.

  • Chinese heavy equipment giants ride AI wave at expo

    Chinese heavy equipment giants ride AI wave at expo

    At the recently concluded CONEXPO-CON/AGG, North America’s premier construction equipment trade show in Las Vegas, Chinese heavy machinery manufacturers demonstrated groundbreaking advancements in artificial intelligence and automation that are reshaping the global industry landscape.

    The event, held from March 3-7, became a showcase for how AI and robotics are transforming traditional construction equipment. Industry experts highlighted that automation addresses critical challenges including labor shortages and operational hazards while delivering substantial efficiency improvements. According to technology consultant Nick Espinosa of Security Fanatics, these technologies can boost productivity by nearly 15% while reducing costs by approximately 12%.

    Chinese manufacturer SANY captured significant attention with its demonstration of remotely operated excavation technology. Visitors experienced operating an excavator located in China from Las Vegas using an “intelligent cab” with joystick controls and live camera feeds across 10,000 kilometers. Jeremiah Percival of PAF Excavation & Dredging, who tested the system, noted: “It was very cool. With minimal familiarization time, operators could become proficient with the entire setup.”

    Beyond remote operation, SANY displayed multiple AI-powered innovations including grade control systems, fleet management platforms, and AI service assistants designed to enhance equipment maintenance efficiency.

    The emergence of Chinese manufacturers represents a strategic shift in the industry historically dominated by century-old Western brands like Caterpillar, John Deere, and Case IH. Rather than directly competing with established players, Chinese companies are leveraging their technological agility in areas including industrial internet, artificial intelligence, and smart manufacturing.

    As Vivin Hegde, founding partner of Zacua Ventures, observed during the conference: “Capital follows problems, and the most pressing problem today is labor shortages. The industry is looking at automation in both software and hardware.”

    Feng Ruoyu, XCMG’s senior brand manager, acknowledged the competitive landscape while emphasizing China’s different approach: “We talk about overtaking on a different track. Our advantages lie in areas like AI and intelligent manufacturing. Being younger gives us more room to make bold changes without legacy burdens.”

  • China leads global AI development with its open-weight models

    China leads global AI development with its open-weight models

    At the forefront of artificial intelligence innovation, China is reshaping global technological collaboration through its pioneering open-weight model approach. During the ongoing National Two Sessions, prominent AI scholar Wang Jian articulated a vision that transcends simplistic US-China competition narratives, emphasizing instead which nation can deliver greater value to humanity through artificial intelligence.

    As a member of the National Committee of the Chinese People’s Political Consultative Conference and Chinese Academy of Engineering academician, Wang contends that China’s open-weight methodology represents a profound commitment to global AI advancement. Unlike conventional open-source practices that primarily share code, open-weight models provide access to the computationally intensive weights behind large language models—effectively sharing the substantial computing power, electricity, and resources already invested.

    “When China makes its large AI models open and transforms them into open-weight models, the significance extends beyond traditional open-source logic,” explained Wang, who directs Zhejiang Lab and founded Alibaba’s cloud computing division. “We are essentially defining how this new developmental phase should progress.”

    This strategic direction aligns with China’s governmental priorities, as reflected in the draft Government Work Report presented to the National People’s Congress, which explicitly supports developing open-source AI communities and cultivating vibrant open-source ecosystems.

    Empirical evidence demonstrates the global impact of China’s approach. A collaborative study between MIT and Hugging Face revealed that Chinese open-source models achieved a 17.1% global download share between August 2024-2025, surpassing the United States’ 15.8% for the first time. DeepSeek and Alibaba’s Qwen emerged as particularly influential contributors to this milestone.

    Microsoft’s January report further documented how Chinese startup DeepSeek has accelerated AI adoption across developing nations through free, open-source models. In multiple African countries including Ethiopia, Zimbabwe, Uganda, and Niger, DeepSeek captured market shares between 11-14%, demonstrating technology’s democratizing potential.

    Wang emphasized that China’s AI development philosophy extends beyond technological prowess to encompass broader societal advancement. The government work report specifically highlights implementing effective measures to facilitate employment and entrepreneurship through AI adaptation.

    “This perspective transcends AI technologies themselves and their immediate applications,” Wang noted. “We place greater emphasis on AI’s capacity to drive comprehensive social development.” He drew parallels with historical innovations like paper and electricity, which generated entirely new industries and professions through creative implementation.

    As China’s scientific capabilities reach world-leading levels, Wang anticipates benefits will extend beyond national borders, particularly assisting Global South and developing nations in their technological transformation journeys.

  • Drone-assisted snow and ice clearing takes off

    Drone-assisted snow and ice clearing takes off

    Across China’s snow-prone regions, a technological transformation is underway as drones take center stage in winter maintenance operations. In remote mountainous areas of Xinjiang Uygur Autonomous Region, where heavy snowfall regularly threatens power infrastructure, utility companies have deployed unmanned aircraft systems to perform critical de-icing tasks that previously endangered human workers.

    The innovative approach involves drones equipped with specialized insulating tools that mechanically dislodge thick ice accumulations from high-voltage power lines. This technological intervention represents a significant departure from traditional methods that required linemen to undertake dangerous climbs in freezing conditions. State Grid Xinjiang Electric Power Company, which implemented heavy-lift drones for de-icing operations starting December 2024, reports substantial operational improvements and cost savings exceeding 1 million yuan ($144,000).

    Beyond power infrastructure, agricultural applications have emerged as another frontier for drone technology. In Hubei province’s Badong county, orange growers like Zhou Zongbao have adopted drone-assisted snow removal techniques. Using rotor-generated airflow, farmers can efficiently clear snow from fruit trees without damaging branches or compromising fruit quality. This method proves particularly valuable for protecting citrus crops where accumulated snow can break branches and cold damage affects both sweetness and appearance.

    The economic impact extends beyond operational efficiency. Drone-based snow removal services have created new business opportunities and employment pathways across multiple provinces. In Suzhou city, Anhui province, companies charge approximately 200 yuan for greenhouse snow clearance, while in Xuchang city, Henan province, certified drone pilots command rates of 300 yuan per hour. Regional variations in pricing reflect factors including job complexity and operational difficulty, but overall returns remain attractive for service providers.

    With over 1,600 drones deployed and more than 1,300 employees certified as drone pilots, State Grid Xinjiang exemplifies the scaling of this technology. The company emphasizes that drone operation has become an essential skill within its workforce, signaling a permanent shift in maintenance methodologies. As winter conditions persist across northern China, drone-assisted ice and snow clearing continues to demonstrate how technological innovation translates into practical solutions for both industrial and agricultural challenges.

  • Regional hub sets sights on world-class innovation

    Regional hub sets sights on world-class innovation

    The Beijing-Tianjin-Hebei economic cluster is accelerating its transformation into a globally competitive innovation corridor, with the Xiong’an New Area emerging as a pivotal hub for advanced aerospace manufacturing. The region’s technological ambitions were demonstrated when Xiong’an No 1, the area’s inaugural domestically produced satellite, completed production in October 2025 at Land-Space HongQing (Xiong’an) Space Technology Co’s state-of-the-art facility.

    Within the company’s satellite assembly plant, automated robotic arms execute precision maneuvers along production lines, handling components with millimeter accuracy. This technical achievement represents more than manufacturing prowess—it signifies Xiong’an’s rapid evolution from conceptual development to tangible high-tech product realization.

    Land-Space HongQing achieved a remarkable industrial milestone by completing company registration, facility construction, production setup, and satellite rollout within a single calendar year. “The velocity of this project exceeds national standards for such complex undertakings,” stated General Manager Shi Yaozhong.

    The company exemplifies a innovative collaborative model between Beijing and Xiong’an. While Beijing headquarters manages administrative functions and research development, the Xiong’an subsidiary specializes in pilot production—the critical phase where engineers refine manufacturing processes and validate technologies under operational conditions.

    Shi emphasized Xiong’an’s strategic advantages: “The area’s prioritization of satellite internet industry development creates an ecosystem conducive to innovation. Proximity to supply chain partners and end-users enables immediate feedback integration during pilot production, facilitating continuous product enhancement.”

    This development aligns with broader regional efforts to establish the Beijing-Tianjin-Hebei cluster as a dominant force in China’s high-technology landscape, potentially reshaping global competitiveness in aerospace and satellite technology sectors.

  • AI-enhanced images of real events distort view of Mideast war

    AI-enhanced images of real events distort view of Mideast war

    The ongoing Middle East conflict has become a testing ground for sophisticated AI-driven disinformation techniques that are fundamentally altering how warfare is perceived. Beyond completely fabricated content, a new category of manipulated imagery is emerging: authentic photographs that have been artificially enhanced to distort reality while maintaining a veneer of credibility.

    A compelling case involves a widely circulated photograph depicting a kneeling US pilot confronted by a Kuwaiti local after parachuting from his aircraft. While the incident itself was verified through satellite imagery and video evidence, AI detection tools revealed the presence of Google’s SynthID watermark, indicating artificial enhancement. The most telling detail? The pilot appears with only four fingers on each hand—a common AI generation artifact.

    According to Professor Evangelos Kanoulas, an AI expert at the University of Amsterdam, these enhancements subtly manipulate textures, facial expressions, lighting conditions, and background details. “This can strengthen a particular narrative about an event—for example, making a protest appear more violent, making a crowd appear larger, or making facial expressions more intense,” he explained.

    Another verified example shows dramatic imagery from Erbil International Airport in Iraq following Iranian strikes on March 1. While the original photograph captured a genuine event, the AI-enhanced version exaggerated the scale of the fire and smoke while intensifying colors to create a more dramatic effect.

    The boundary between enhancement and outright fabrication is becoming increasingly blurred. Professor James O’Brien of UC Berkeley’s Computer Science Department warned that “even little changes can end up telling a very different story and could change the perception of events.”

    This phenomenon manifested disturbingly following the shooting of Alex Pretti in Minneapolis, where an AI-enhanced image based on genuine footage incorrectly transformed a phone in the victim’s hand into what appeared to be a weapon—fundamentally altering the narrative of the event.

    As the Middle East conflict continues, experts express grave concerns about the erosion of public trust. Without proper labeling standards, AI-enhanced imagery not only spreads misinformation but also creates widespread skepticism toward authentic documentation. This dual effect—amplifying false narratives while undermining genuine evidence—represents a fundamental challenge to truth verification in modern conflict reporting.