分类： technology

The landmark partnership between TikTok and Oracle Corporation continues to generate uncertainty regarding the platform’s operational future for its American user base. While the arrangement nominally addresses national security concerns raised by U.S. regulators, industry observers note the conspicuous absence of detailed specifications governing user experience modifications.

This strategic alliance, formed under intense geopolitical pressure, ostensibly creates a new entity called TikTok Global to oversee U.S. operations. Oracle’s role as a “trusted technology partner” provides the architectural framework for data security compliance, yet the practical implications for content delivery algorithms, data processing protocols, and feature accessibility remain deliberately vague in public disclosures.

The arrangement’s ambiguity stems from complex negotiations involving multiple stakeholders: ByteDance (TikTok’s Chinese parent company), U.S. regulatory bodies, and now Oracle as the primary American technology partner. This tripartite dynamic creates inherent tensions between national security priorities, commercial interests, and user experience consistency.

Technology analysts suggest the partnership could potentially introduce differentiated user experiences across geographical regions, with American users possibly encountering modified content recommendation algorithms or altered data retention policies. The fundamental architecture of TikTok’s service—particularly its critically acclaimed content discovery mechanism—may undergo significant recalibration to satisfy regulatory requirements.

Market response indicates cautious optimism tempered by practical concerns. While the partnership prevents immediate platform prohibition in the United States, preserving access for TikTok’s approximately 100 million American users, the long-term user experience implications warrant continued scrutiny as implementation details emerge.

BANGKOK — In a landmark move against transnational cybercrime, Thailand has spearheaded the formation of an international coalition to tackle the escalating threat of online scams that cost global victims an estimated $18-37 billion annually. The initiative, formally launched as the Global Partnership Against Online Scams, emerged from a two-day conference co-hosted by Thailand’s Ministry of Foreign Affairs and the United Nations Office on Drugs and Crime (UNODC).

The gathering brought together over 300 participants from nearly 60 nations, culminating in a multilateral agreement signed by Thailand, Bangladesh, Nepal, Peru, and the United Arab Emirates. The partnership framework encompasses four critical pillars: political commitment, enhanced law enforcement cooperation, comprehensive victim protection protocols, and cross-border public awareness campaigns.

Thai Prime Minister Anutin Charnvirakul set the tone in his keynote address, emphasizing that ‘online scams reveal a deeper problem—a collective vulnerability that no country can address alone.’ This sentiment echoed throughout the conference, which highlighted how criminal enterprises, particularly those operating across Southeast Asia, have evolved to exploit digital platforms and artificial intelligence technologies.

The private sector played a significant role in the proceedings, with tech giants Meta and TikTok contributing expertise and resources. Meta presented a threat assessment detailing how scam networks are increasingly leveraging AI to perpetrate fraud across its platforms including Facebook, Instagram, and WhatsApp. TikTok formally endorsed the conference’s closing statement, becoming one of the first corporate members of the new alliance.

Brian Hanley, Asia-Pacific director of the Global Anti-Scam Alliance (which TikTok recently joined), stressed the necessity of multi-stakeholder engagement: ‘Scams are exploiting not only transnational boundaries but also the seams across various platforms from banks, telcos, to social media platforms.’ He emphasized that combating these sophisticated networks requires ‘all major stakeholders at the table.’

The urgency of the initiative has been amplified by recent developments across the region, including scam center raids in Myanmar, victim repatriation challenges in Thailand, and the tragic death of a South Korean student forced into scam operations in Cambodia. While Cambodia—known as a hub for scam compounds—was notably absent from the conference due to ongoing tensions with Thailand, the collaboration represents growing regional consensus on addressing cybercrime.

This new partnership builds upon previous multilateral efforts, including the United Nations Convention against Cybercrime signed by over 70 countries in Vietnam last October. As UN Secretary-General António Guterres characterized that agreement, such initiatives represent ‘a vow that no country, no matter their level of development, will be left defenseless against cybercrime.’

China has officially commissioned a groundbreaking national network infrastructure that demonstrates unprecedented data transmission capabilities, marking a significant advancement in global information technology. The China Environment for Network Innovation (CENI), the country’s first national science and technology infrastructure in the ICT sector, recently passed national acceptance in Nanjing and is now operational.

This revolutionary network testbed achieved what previously seemed impossible: transferring 72 terabytes of data generated by China’s Five-hundred-meter Aperture Spherical radio Telescope (FAST) from Guizhou province to Huazhong University of Science and Technology in Hubei province in merely 1.6 hours. Traditional internet methods would have required approximately 699 days to accomplish the same data transfer.

According to Liu Yunjie, academician of the Chinese Academy of Engineering, CENI represents China’s entry into the world’s advanced echelon in network technology innovation. “CENI’s passage of national acceptance enables China to deliver higher-quality services for all industries to embrace AI,” Liu stated.

The infrastructure boasts large-scale, programmable, virtualized, and open-sharing features, providing a simple, efficient, and low-cost test environment for researching innovative architectures of future networks. The development team has pioneered a service-customized network architecture internationally and created the world’s first distributed large-network operating system.

After over a decade of construction, CENI now spans 40 cities nationwide with a total optical transmission length exceeding 55,000 kilometers. The network can support 128 heterogeneous networks and 4,096 parallel heterogeneous service tests while seamlessly interconnecting with existing internet infrastructure and global test facilities.

The project has generated substantial intellectual property, including 206 international and domestic standards, 221 authorized invention patents, and 139 software copyrights. Multiple achievements have received national-level scientific and technological awards.

Wu Hequan, another academician of the Chinese Academy of Engineering, emphasized that CENI will support the development of 5G-Advanced and 6G technologies while allowing research institutions and enterprises to test various new technologies.

Notably, CENI offers dramatic improvements for AI development. Liu explained that “training a large model with 100-billion-level parameters across domains requires over 500,000 rounds of iterative development. With CENI, each training round takes only about 16 seconds,” resulting in significant time and economic savings.

Looking forward, CENI will provide open testbed support to key industries including manufacturing, energy and power, education, healthcare, and the low-altitude economy, thereby accelerating China’s digital transformation initiatives.

In a landmark advancement for neurotechnology, Chinese researchers have demonstrated unprecedented brain-machine interface (BMI) capabilities enabling paralyzed patients to interact with their physical environment through thought alone. Two individuals with high-level paralysis have successfully operated power wheelchairs, directed robotic dogs to retrieve objects, and manipulated robotic arms to perform tasks like drinking water—all without physical movement.

The achievement, announced by the Center for Excellence in Brain Science and Intelligence Technology of the Chinese Academy of Sciences, represents a significant leap beyond previous BMI applications that were limited to virtual interactions. The patients, both in their 30s, received minimally invasive neural implants developed through collaboration between the CAS center, Shanghai Huashan Hospital, and industry partners.

Technical innovations were crucial to this breakthrough. Researchers overcame substantial challenges including neural data compression, environmental noise interference, and maintaining ‘cross-day stability’ through neural manifold alignment techniques. Most impressively, the team achieved an end-to-end delay of under 100 milliseconds from neural signal to device execution—faster than the body’s natural neural transmission speed.

The system’s practicality is enhanced by its cost-effective design and minimal invasiveness. The implant requires only a 5-millimeter cranial puncture, the smallest reported globally, and is approximately half the size of Neuralink’s device. According to lead scientist Zhao Zhengtuo, the technology is progressing toward clinical applications with the goal of making implantation as simple as ear piercing.

Future development timelines anticipate motor and language function restoration within three years, sensory restoration and treatment for neurological disorders within five years, and highly minimally invasive systems for both medical and consumer applications within a decade. All procedures underwent rigorous ethical and technical reviews, with patients reported in good health throughout the trials.

In a landmark advancement for neurotechnology, Chinese researchers have demonstrated unprecedented brain-machine interface (BMI) capabilities enabling two paralyzed patients to perform complex physical tasks through thought alone. The breakthrough achievements—announced December 17, 2025, by the Center for Excellence in Brain Science and Intelligence Technology of the Chinese Academy of Sciences—include the world’s first mental control of a power wheelchair and robotic dog for delivery retrieval, plus precise manipulation of a robotic arm for drinking activities.

The two patients, both males in their thirties suffering from high-level paralysis, received minimally invasive BMI implants developed through collaboration between the CAS center, Shanghai Huashan Hospital, and corporate partners. The implant procedures—conducted in June and October 2025 through a mere 5-millimeter cranial puncture—represent the smallest such implementation globally, with the implant device itself being approximately half the size of Neuralink’s comparable technology.

This research signifies a quantum leap beyond the team’s previous March 2025 achievement where a patient mentally controlled a computer cursor. The new cases enable three-dimensional interaction with the physical environment, dramatically expanding possibilities for self-care, employment, and social participation. One patient has already secured employment as an intern product sorter, utilizing brain control to perform online data annotation for AI verification systems in vending machines.

Technical innovations were crucial to these advancements. Researchers developed high-compression, high-fidelity neural data compression techniques and created hybrid decoding models capable of extracting usable signals in noisy environments. By addressing ‘cross-day stability’ challenges through neural manifold alignment, the team maintained reliable decoder performance despite environmental interference and physiological variations.

Perhaps most impressively, the system achieves end-to-end delay of under 100 milliseconds from neural signal acquisition to device execution—faster than the body’s natural neural transmission speed. Patients describe the experience as intuitive and seamless, comparable to controlling video game characters without conscious effort.

According to lead scientist Zhao Zhengtuo, the research represents significant progress toward practical clinical applications. The team anticipates scalable applications for restoring motor and language functions within three years, breakthroughs in sensory restoration and neuropsychiatric disorder treatment within five years, and highly minimally invasive systems enabling both medical and consumer uses within a decade. Future goals include achieving sufficiently fine control for activities such as playing the piano with mind-controlled robotic fingers.

In an unprecedented strategic maneuver, global AI giants OpenAI, Google, and Perplexity have launched an aggressive campaign to capture India’s massive user base through complimentary premium subscriptions. This calculated initiative represents more than mere user acquisition—it constitutes a crucial effort to gather diverse multilingual training data from the world’s most populous nation.

Market intelligence from Sensor Tower reveals staggering adoption metrics: ChatGPT now commands 73 million daily users in India, more than double its U.S. user base, following a remarkable 607% year-on-year surge. Similarly, Google’s Gemini has experienced a 15% growth spike since November, reaching 17 million Indian daily users compared to just 3 million in the United States.

The strategic free offerings are substantial: Google provides its $400 Gemini AI Pro subscription at no cost for 18 months to 500 million Reliance Jio subscribers. OpenAI has made its ChatGPT Go plan completely free for one year across India, previously priced at $54 locally. Perplexity has similarly waived its $200 annual Pro subscription fee for Airtel telecommunications users.

Behind these generous offers lies a critical data acquisition strategy. Five AI analysts confirm that India’s linguistic diversity presents a unique training opportunity for AI models. The complex communication patterns, multiple languages, and distinctive dialects characteristic of Indian users provide invaluable data that current training sets lack.

Sagar Vishnoi, co-founder of Future Shift Labs, explains: ‘These complimentary initiatives effectively address gaps in AI training datasets that currently lack comprehensive information on regional user behavior patterns.’

The approach leverages India’s exceptional digital landscape: 730 million smartphone users consuming 21 gigabytes of data monthly at just 9.2 cents per gigabyte—among the world’s most affordable mobile data rates. This price sensitivity has historically responded well to free access models, as demonstrated by Reliance’s successful customer acquisition strategies.

While users like Hyderabad PhD student Anees Hassan enthusiastically utilize these tools for research and presentation tasks, concerns regarding data harvesting practices persist. Hassan acknowledges implementing opt-out features to prevent his data from contributing to AI training, highlighting the privacy considerations accompanying these free access programs.