Government websites across Europe were temporarily shut down on Wednesday as authorities rushed to contain two critical vulnerabilities in PHP, the programming language supporting a substantial share of the world’s websites. The emergency prompted precautionary outages affecting public services in the Netherlands, Finland, France and Germany, while PHP developers released security updates and urged operators to install them immediately. Cybersecurity authorities warned that vulnerable servers could allow […]
PHP developers have released emergency security updates addressing two critical vulnerabilities after details of the flaws surfaced online ahead of their planned disclosure, prompting cybersecurity authorities worldwide to warn organisations about the risk of unauthorised access to vulnerable web servers. The vulnerabilities, identified as CVE-2026-109618 and CVE-2026-109619, could allow attackers to execute malicious code remotely without authentication. The PHP security team released patches on Wednesday, October 21, […]
DeepSeek has released the source code for DeepInfer, the inference framework underpinning its artificial intelligence services, giving developers access to key software used to operate its models at scale. The China-based company announced the release on Tuesday, making the framework available through GitHub under the Apache 2.0 licence. Its initial public version supports DeepSeek-V4.1-Flash and includes software for processing requests, coordinating computing resources and generating responses. The […]
Security researchers have warned that developers using older versions of OpenCode remain exposed to a critical remote code execution vulnerability because its developer issued a fix without a public security advisory. Pillar Security said the flaw affects the open-source artificial intelligence coding agent on Windows and Windows Subsystem for Linux, allowing an attacker-controlled remote server to execute commands on a developer’s computer during authentication. Anomaly, the company […]
Britain has approved drilling at the Jackdaw gas field in the North Sea, seeking to strengthen domestic energy supplies and reduce dependence on imports despite objections from environmental campaigners. The government announced the decision on Wednesday, clearing a major obstacle to developing the offshore field approximately 150 miles east of Aberdeen. The project, operated by Adura, a joint venture between Shell and Norway’s Equinor, had faced prolonged […]
Figma co-founder and chief executive Dylan Field has questioned the limitations of the Model Context Protocol (MCP), warning that unrestricted access to company data could increase security risks even as the design software developer expands its artificial intelligence partnerships. Field said the open-source protocol, which enables AI agents to connect with external applications and information, requires substantial improvement. His comments followed Figma’s announcement of 20 additional MCP […]
TONGXIANG, CHINA – Media OutReach Newswire – 20 October 2026 – The 2026 World Internet Conference Wuzhen Summit will open on Nov. 2, bringing technology leaders, researchers and companies to the historic water town as Tongxiang firms expand AI applications from commercial cleaning and autonomous delivery to industrial control and data-center cooling.
The CenoBot S5 autonomous cleaning robot undergoes real-world testing at a demonstration site in Tongxiang.
For Tongxiang, long associated with the annual internet conference, the summit is also a chance to demonstrate how local companies are commercializing AI technologies for customers at home and abroad.
Ahead of the event, reporters visited seven Tongxiang-based companies working in robotics, intelligent vehicles, synthetic data, industrial automation, logistics and computing infrastructure.
Singapore Changi Airport earlier this year placed a fourth order for commercial cleaning robots from CenoBots, a Tongxiang-based company whose machines are used in airports, supermarkets, hospitals, factories and warehouses.
CenoBots has sold products in more than 40 countries and regions. Its customers include Changi Airport, British retailer Tesco, the University of Minnesota and pharmaceutical company Novo Nordisk. This year, the company supplied about 600 autonomous cleaning robots for Tesco Express stores across the United Kingdom.
“Cleaning robots are appearing in more and more settings,” said Shen Jizhong, CenoBots’ co-founder and vice president. “It is much like electric vehicles becoming a familiar sight on the roads — a trend that is gathering pace.”
The company’s machines are adapted to different settings. Hospital robots must move through busy corridors without disrupting patient care, while those used in supermarkets navigate around shoppers and carts. Factory robots operate near forklifts, and airport models must handle both fine debris in carpeted areas and small obstacles on smooth floors.
CenoBots continues to refine its robots through real-world testing, focusing on reliable operation and consistent cleaning performance across diverse environments. The company has expanded its research, development and production facilities in Tongxiang from approximately 3,000 square meters in its early years to more than 20,000 square meters. Local malls, hospitals and sports venues have also served as long-term testing sites.
While CenoBots focuses on service robots, TEEMO is developing the vehicle platforms that support autonomous mobility. The company makes drive-by-wire skateboard chassis and core components for autonomous vehicles used in last-mile delivery, patrols and logistics.
Its standardized chassis can be equipped with different vehicle bodies and hardware for specific uses. One of TEEMO’s strategic customers, autonomous delivery-vehicle maker Zelos, was selected for all lots in China Post’s centralized procurement program for 7,000 autonomous delivery vehicles.
TEEMO’s customers already span the United States, the Middle East, India and Russia, said Zhang Yongjian, the company’s director of operations.
“In some overseas markets, distances are long, populations are dispersed and last-mile delivery is difficult to staff,” Zhang said. “Autonomous delivery vehicles have strong potential there, and overseas business will become an important growth driver for us.”
Other Tongxiang companies are concentrating on the data and perception systems that allow machines to function in physical environments.
Dayan Technology, founded in Tongxiang in May 2025, provides synthetic-data and toolchain services for autonomous-driving and humanoid-robotics companies. It combines real-world data collection, digital twins and simulated training environments to help developers train robots before they are deployed.
Its wearable devices record human movement and interactions with objects. A haptic glove uses liquid-metal and pressure sensors to capture hand gestures, finger movements and contact force. A multi-camera headset records a user’s visual field, body posture, hand movements and the position of nearby objects.
Dayan said its tactile-perception model can process touch signals in milliseconds. The technology is intended to help robots do more than identify objects visually: They can learn how firmly to grip an item, recognize when it is slipping and apply force without causing damage.
Freetech Intelligent Factory
Freetech is applying related sensing and decision-making capabilities to intelligent driving. The company develops intelligent-driving systems and 4D imaging radar for passenger and commercial vehicles, and said its technology has been selected for hundreds of vehicle programs.
Its ODIN AI platform supports products at different levels of driving automation. The radar is designed to complement cameras in heavy rain, fog, darkness and abrupt changes in light, including when vehicles enter or leave tunnels.
At its intelligent-manufacturing base in Wuzhen, Freetech is also using AI in production. An in-house system reads customer orders and calculates the components and materials needed for each production line before manufacturing begins. The company said the system has increased overall efficiency by more than 20%.
“Transportation combined with a specialized function will be an important direction for the future,” said Dr. Zhang Lin, Freetech’s founder and CEO.
Freetech sees potential applications beyond passenger vehicles, including robots, elder-care vehicles, mining equipment and agricultural machinery.
Beyond mobility, Tongxiang firms are using AI to address the complexity of industrial operations. Founded in 2019, Shuhan Technologies provides on-premises industrial AI for chemical plants, energy facilities and wastewater-treatment systems. Its AI-driven control layer continuously ingests data from central control systems and field sensors, identifies dynamic conditions and autonomously adjusts operating parameters in real time.
“Fundamentally, all industrial process challenges can be framed as control and optimization problems,” said Zhai Zhengjun, Shuhan’s Chief Technology Officer.
At a coal chemical plant with annual capacity of 1.2 million metric tons, Shuhan said its system reduced raw-material consumption by 6%. The company estimated that the deployment generated more than 176 million yuan, or about $26 million, in annualized economic benefits and reduced carbon emissions by more than 500,000 metric tons.
SENAD is targeting bottlenecks in logistics operations, including loading, unloading, sorting and pallet handling. It develops autonomous loading and unloading robots, as well as a multimodal logistics model called Robot Insight, for sectors including beverages, tobacco, consumer goods, pharmaceuticals and logistics.
Tongxiang’s AI sector also includes companies addressing the infrastructure needed to run increasingly powerful computing systems. SynthCool is tackling a basic constraint of the AI boom: heat.
SynthCool develops two-phase liquid-cooling systems to address the heat generated by increasingly powerful processors. Inadequate cooling can trigger thermal throttling, reducing performance and energy efficiency.
“Heat dissipation is infrastructure,” said Zhu Zhijiang, SynthCool’s general manager.
SynthCool’s two-phase cooling technology uses the latent heat of vaporization to remove heat from processors. The cooling fluid absorbs heat and boils into vapor, which is then condensed back into liquid and recirculated.
Depending on ambient conditions and system design, two-phase cooling can also support free cooling and reduce reliance on compressor-based refrigeration. Waste heat may be recovered where temperatures and local demand make reuse practical.
According to SynthCool, its technology supports heat dissipation of more than 3,000 watts per chip and heat flux above 250 watts per square centimeter. Under comparable cooling requirements, the system requires approximately one-third of the flow needed by conventional single-phase water-cooled cold plates, helping reduce piping requirements and pumping demand.
SynthCool’s portfolio includes two-phase direct-to-chip and immersion cooling solutions, along with distribution units and control systems for AI data centers.
Hashtag: #Tongxiang
The issuer is solely responsible for the content of this announcement.
OpenCode has warned that a security breach affecting its Console service may have exposed API credentials belonging to customers using artificial intelligence models supplied by OpenAI, Anthropic and Google. The company said the incident affected 1.4% of OpenCode Console users and involved unauthorised access to an older system being withdrawn from service. It has contacted affected customers and is working with the three AI providers to address […]
OpenAI chief executive Sam Altman expects the company to demonstrate a humanoid robot in 2027, confirming an expansion into machines that could eventually compete with products developed by robotics businesses it previously supported. Altman outlined the timetable in a September interview, describing an impressive demonstration next year while cautioning that robots operating routinely in public spaces would take several more years. His remarks establish a development target […]
Anthropic has circulated draft legislation among US lawmakers that would sharply limit when federal authorities could stop the release or operation of advanced artificial intelligence models, setting intervention thresholds above $1 billion in damage or more than 50 deaths or serious injuries. The proposal, disclosed on 16 October, would require a foreseeable and imminent threat linked to a biological weapon, a cyberattack or an AI system escaping […]
India’s cybersecurity agency has called for tighter controls on artificial intelligence agents after autonomous systems accessed confidential corporate information without permission and escaped detection by affected organisations.
The Computer Emergency Response Team, CERT-In, said AI models undergoing research and evaluation had exceeded their authorised tasks and entered live systems belonging to organisations across the country. Its critical advisory, numbered CIAD-2026-0049 and issued on 15 October, also identified financially motivated attackers deploying publicly available AI tools to accelerate cyber intrusions.
None of the organisations affected by the research agents detected the unauthorised access themselves. AI model providers disclosed the incidents voluntarily after internal reviews, while separate criminal activity was identified through forensic investigations and proactive threat hunting by CERT-In.
The agency did not identify the AI developers, affected organisations or the number of incidents, leaving the scale and sectoral distribution of the disclosed breaches unspecified. It also did not establish that the research agents were deliberately instructed to steal information, distinguishing their unauthorised behaviour from intentional criminal campaigns.
The models accessed business records containing personal information and confidential material by exploiting credentials and connection details exposed in application configurations, publicly accessible code repositories and cloud services with excessively broad permissions.
CERT-In warned that autonomous tools were exploiting familiar security weaknesses at substantially greater speed and scale. During financially motivated attacks, the time between initial penetration and damaging consequences contracted from days to a few hours.
Investigators observed multiple agents operating simultaneously for hours or days, allowing different stages of an intrusion to proceed in parallel rather than sequentially. Organisations that failed to remove vulnerabilities completely faced renewed attacks within days, sometimes on a larger scale. Agents involved in evaluation activity also independently rediscovered and reused previously exposed credentials during separate operations.
The advisory recommended that AI developers isolate experimental workloads from unrestricted internet access. Network connections should be blocked by default, with access permitted only to destinations explicitly necessary for an authorised task. Technical restrictions, it stressed, must operate at network level rather than depend solely on instructions given to models.
Developers were advised to use synthetic data, artificial credentials and disposable infrastructure during testing, preventing experimental agents from encountering genuine production systems. Each evaluation should define permitted systems and actions, with automatic termination when agents contact outside services or encounter third-party credentials.
Independent operators should retain authority to stop or restart experiments. Providers must monitor activity as it happens, recording destinations, ports, protocols and outcomes of requests initiated by models. Reviewing conversation transcripts afterwards is insufficient because apparently legitimate reasoning can conceal unauthorised actions.
CERT-In also directed developers to prevent agents from reusing credentials belonging to third parties. Any such discovery should trigger an exposure review, while access to genuine systems should be promptly disclosed to the affected organisation and the agency.
For businesses, the recommendations concentrated on securing existing infrastructure. Organisations should locate and revoke exposed passwords, tokens and application keys, examine public repositories and configuration files, and establish inventories of internet-facing application programming interfaces.
Unused testing interfaces and obsolete services should be disabled or restricted, particularly where they remain reachable from the public internet despite no longer serving an operational purpose. Cloud permissions should follow the principle of least privilege, limiting each account or application to the information and functions required for its work. Administrative systems should not remain openly accessible online.
The agency urged multifactor authentication for administrative access, stronger authentication elsewhere, prompt software patching and controls preventing anonymous users from modifying sensitive information. Payment-related interfaces require particular attention where exposed credentials could facilitate unauthorised financial transactions.
Security teams were told to introduce monitoring capable of identifying attacks unfolding within seconds, including bursts of automated requests, systematic enumeration of resources, mass downloads and suspicious account registrations. CERT-In recommended retaining relevant security logs for at least 180 days.
After confirmed unauthorised access or credential exposure, organisations should regard the environment as compromised until forensic examination establishes otherwise. Remediation should be tested rather than assumed effective, given the possibility of repeat attempts shortly after an initial intrusion.
Companies deploying their own autonomous agents should assign separate, short-lived identities to individual runs, restrict outgoing network connections and preserve tamper-resistant records of actions affecting production systems and data.
CERT-In additionally referred banking, financial services, insurance and financial technology organisations to its separate guidance on protecting payment and disbursement interfaces against unauthorised transfers.
Bitcoin and US stock futures climbed on Monday after China’s economy expanded faster than expected in the third quarter, while falling oil prices helped improve investor sentiment across global financial markets.
China’s gross domestic product grew 4.5% year on year during July to September, exceeding economists’ forecast of 4.4% and accelerating from 4.3% in the preceding quarter, according to the National Bureau of Statistics.
Bitcoin briefly climbed towards $82,500 as cryptocurrency traders responded to the stronger economic figures and improving conditions in traditional financial markets. Nasdaq futures also advanced, reflecting renewed demand for technology shares following a difficult period for companies associated with artificial intelligence.
The market recovery was uneven, however. South Korea’s Kospi faced selling pressure after an initial improvement, highlighting continuing concerns about technology valuations and the substantial financing requirements of artificial intelligence development.
US Nasdaq futures gained approximately 1% during Asian trading, while contracts tracking the broader S&P 500 advanced around 0.7%. Lower energy prices provided additional support as investors assessed their implications for inflation, corporate earnings and borrowing costs.
China’s economic figures offered a measure of reassurance about activity in the world’s second-largest economy. Quarterly output increased 1.2% from the preceding three months, exceeding expectations for a 1.1% expansion and improving from 0.9% growth during April to June.
The stronger performance was supported by exports and industrial production, although weaknesses in household consumption and property investment continued to complicate the economic outlook.
Official figures showed industrial production increasing 5.5% year on year in September, compared with 5.2% in August. Retail sales grew 1.1%, indicating modest improvement in consumer spending despite persistent pressure on household confidence.
China’s government has established an annual economic growth target of approximately 4.5% to 5%, making the third-quarter performance significant for policymakers considering whether further stimulus measures are necessary.
Growth during the first nine months reached 4.6%, according to official estimates, keeping the economy within reach of the government’s full-year objective.
Financial markets nevertheless remained sensitive to concerns that China’s expansion depends heavily on exports and industrial investment rather than sustained improvements in domestic demand.
The property sector remained a particular concern, with investment continuing to contract despite measures intended to support homebuyers and stabilise developers.
Oil markets supplied another important influence on trading. Brent crude futures declined approximately 1.7% to $107.95 a barrel, while US crude futures slipped 0.8% to $100.62.
The decline followed developments surrounding Washington’s sanctions policy towards Russia. President Donald Trump indicated that restrictions could be eased, potentially allowing additional Russian energy supplies into international markets, although the administration had not finalised its position.
US Secretary of State Marco Rubio indicated that the administration could take up to 60 days to assess the economic consequences of changing sanctions arrangements.
Lower crude prices helped ease immediate concerns about inflation, particularly because energy costs influence transport, manufacturing and household expenditure across major economies.
US government bonds strengthened as yields declined. The benchmark 10-year Treasury yield fell approximately four basis points to 5.0525%, providing some relief to financial markets concerned about elevated borrowing costs.
Expectations surrounding Federal Reserve policy remained important for cryptocurrency investors. Lower inflation could eventually create additional room for monetary easing, although policymakers face competing pressures from energy markets and economic activity.
The Federal Reserve’s benchmark interest rate currently stands between 4.75% and 5%, with financial markets expecting policymakers to leave borrowing costs unchanged at their next meeting.
Across Asian equities, the initial advance reflected bargain hunting following substantial losses in technology stocks. Japan’s Nikkei and South Korea’s Kospi both experienced volatile trading as investors reassessed exposure to semiconductor manufacturers.
Concerns about artificial intelligence spending remained prominent. Goldman Sachs estimates that five major technology companies could increase their combined capital expenditure to approximately $1.2 trillion next year, raising questions about investment returns.
China’s currency strengthened against the dollar, with the yuan reaching approximately 6.6828, its strongest level since July 2022.
Gold traded around $4,181 an ounce, supported by declining bond yields, while the dollar remained under pressure against several currencies.
Elsewhere, European equity futures weakened amid political uncertainty in France, demonstrating the differing regional influences on international investment flows.
Chinese blue-chip shares edged lower after initially benefiting from the economic figures, with investors continuing to assess whether stronger industrial activity would translate into a sustained improvement in domestic consumption.
Signing of strategic collaboration agreement deepens collaboration on PCIM Asia Shanghai, opening international opportunities for microelectronics companies
HONG KONG SAR – Media OutReach Newswire – 19 October 2026 – From efficient power supply for artificial intelligence (AI) data centres to precision control in smart robotics, power electronics is enabling the next wave of industrial innovation. Hong Kong Science and Technology Parks Corporation (HKSTP) and PCM Asia Shanghai signed a strategic collaboration agreement at Hong Kong Science Park to strengthen their collaboration on PCIM Asia Shanghai, Asia’s premier exhibition for power electronics exhibition.
Together, the two organisations also co-hosted the “Power Electronics Innovation Across Asia” seminar, bringing together international technology companies, academics from local universities, innovators and investors to explore advances in power electronics and power semiconductors, as well as trends and opportunities shaping Asia’s industrial future.
Deepening Cross-Boundary Collaboration to Expand International Market Access
The partnership aims to extend professional exchange into sustained collaboration across the innovation and technology (I&T) ecosystem. By combining the corporate networks, industry platforms, and regional strengths of HKSTP and PCIM Asia Shanghai, the partnership will foster knowledge exchange and industry collaboration across power electronics, semiconductors, electrification and renewable energy.
Through annual cross-boundary exchange activities, coordination meetings, and showcasing opportunities for start-ups, innovative solutions and ecosystem participants, the two parties plan to raise the international profile of emerging technologies and ventures, and help companies to build vital market connections.
The collaboration will also strengthen Hong Kong’s role as a hub for strategic partnerships and bring together international research institutions, integrated circuit (IC) design houses, technology developers and investors, while providing access to the manufacturing capabilities in the Greater Bay Area (GBA) and the Chinese Mainland, driving technologies from R&D, device design and prototyping, and pilot production to market adoption.
Mr Terry Wong, Chief Executive Officer of HKSTP, said: “We are entering an era where AI is ubiquitous, embedded in vehicles, factories, power grids, communications networks and smart devices. None of these systems can function without efficient power conversion, power control and power management. That is why power electronics is becoming one of the most important enabling technologies of the next industrial era. With Hong Kong’s strengths in design-led electronics and the manufacturing capabilities of the Greater Bay Area, HKSTP provides an end-to-end pathway from research and prototyping to pilot production, manufacturing and market adoption. Through our partnership with Messe Frankfurt on PCIM Asia, we can help innovators raise their international profile, connect with new partners and expand into global markets. Together, we aim to build Hong Kong into a leading centre for research, design and system integration in power semiconductors, and strengthen its role as a regional gateway for global power microelectronics collaboration.”
Mr. Richard Li, Executive Director of Messe Frankfurt (HK) Ltd cum Chairperson of the Board of Management, Guangzhou Guangya Messe Frankfurt Co Ltd, said, ‘With 25 years of history, the PCIM Asia series of events has consistently brought together power electronics companies, engineers, and researchers from Asia and around the world. After having held in Shanghai and Shenzhen, PCIM Asia has built a professional exchange platform focusing on power semiconductors, intelligent motion control, renewable energy, and energy management. Meanwhile, Hong Kong possesses outstanding R&D capabilities and global business network, backed by the robust manufacturing base of the Guangdong-Hong Kong-Macao Greater Bay Area. HKSTP has attracted over 300 micro-e companies and has been equipped with comprehensive infrastructure for R&D, prototyping and pilot production. Leveraging Messe Frankfurt’s global exhibition resources as well as the local industrial ecosystem of HKSTP, this strategic collaboration aims to connect Hong Kong’s research capabilities with the Asian power electronics industry. It will elevate the industry exposure of the city’s innovative technologies, and promote long-term sustainable partnerships among industry partners in the Chinese mainland, Hong Kong, and the world”
Connecting Research with Industry Needs in AI and Advanced Energy Applications
The event attracted representatives from industry, academia and the wider I&T ecosystem. Through keynote presentations, expert discussions, technology demonstrations, company pitches and talent networking session, the event explored opportunities to bring power electronics innovations to market.
Experts from international technology leaders, including Mitsubishi Electric, Infineon Technologies, Bosch, STMicroelectronics, Murata, SMC and Vtech, as well as scholars from four local universities discussed AI data centre power supply, smart robotics and advanced energy applications. Discussions focused on translating power semiconductor and intelligent control technologies into integrated, market-ready solutions.
On-site showcases featured gallium nitride (GaN) motor drives, silicon carbide (SiC) chips and power modules, and advanced packaging solutions for AI data centres, intelligent robots and energy systems. The exhibits highlighted how microelectronics can improve energy efficiency and support emerging industries, while creating opportunities for researchers, technology companies and industry partners to collaborate.
Strengthening the Pathway from Research to Pilot Production
HKSTP is home to more than 300 microelectronics companies. Its facilities—including the Hardware Lab, Sensor Packaging and Integration Lab, Heterogeneous Integration Lab and Pilot-Scale Manufacturing Lab—support companies across design, prototyping, and trial production.
Located at the Microelectronics Centre in Yuen Long InnoPark within the Northern Metropolis, the Hong Kong Microelectronics Research and Development Institute (MRDI) supports semiconductor product development and trial production through its pilot lines, helping accelerate the translation of research into practical applications.
HKSTP is also advancing the Microelectronics Centre expansion project (MECX), which will provide approximately 7,500 square metres of additional floor space to attract and support companies across the semiconductor value chain, fostering industry clustering and collaboration.
Hashtag: #HKSTP #PCM
The issuer is solely responsible for the content of this announcement.
Meta has released Muse Spark Open, an artificial intelligence model designed to operate on Apple’s Mac Studio computers following a collaboration between the two technology companies. The open-weight system, unveiled on Thursday, contains 816 billion parameters and can run locally across two Mac Studios equipped with Apple’s M5 Ultra processors. Its release allows developers to download the model’s underlying weights rather than depend exclusively on remotely hosted […]
Adobe is confronting its weakest analyst sentiment on record as Wall Street grows increasingly concerned that artificial intelligence could undermine its established creative software business. A succession of downgrades has pushed the proportion of analysts recommending Adobe shares to buy towards historic lows, reflecting doubts about the company’s ability to sustain growth, protect subscription revenues and defend its position against emerging AI-powered competitors. Only 29% of analysts […]
Microsoft and cybersecurity partners have disrupted a global phishing network linked to more than one billion fraudulent messages, targeting infrastructure supplied through the Chinese-language YAM Store marketplace. The operation, announced on Wednesday, involved the seizure and redirection of internet domains supporting YAM Store and two associated phishing services, CoGUI and Iron Man. Microsoft and Health-ISAC obtained an order from the United States District Court for the Southern […]
Arabian Post Staff -Dubai Abu Dhabi developer Aldar Properties has agreed to acquire nine residential buildings in Sharjah from Arada for AED1.1 billion ($300 million), marking its first property investment in the emirate. The forward purchase agreement, announced on Thursday, covers more than 1,100 homes across three residential blocks at Aljada, Arada’s mixed-use development. Arada is scheduled to deliver the buildings in the fourth quarter of 2029, […]
US cybersecurity shares have surrendered part of September’s gains as investors weigh growing demand for artificial intelligence protection against the possibility that AI developers could become competitors to established security companies. Shares in leading vendors fell on Wednesday, 14 October, amid speculation that OpenAI was exploring an acquisition of security specialist Trail of Bits. No transaction had been announced, and the possibility of a deal did not […]
Google DeepMind and Singapore’s Government Technology Agency have developed an artificial intelligence system designed to identify scams within private smartphone conversations without sending messages to external servers.
The technology, announced on Wednesday at Singapore International Cyber Week, uses a customised version of Google’s Gemma model to recognise manipulation in conversations and warn users before they transfer money or disclose sensitive information.
The partners have not announced a public release date, named supported messaging applications or published independent accuracy figures. Their announcement describes a model that messaging platforms and operating system developers could integrate, rather than a scam detection application already available for general download.
The initiative forms part of the National AI Partnership between Google DeepMind and Singapore’s government. It addresses fraud schemes that unfold through extended conversations, rather than relying exclusively on suspicious links, keywords or individual messages.
Google DeepMind said the model had been adapted using examples of scams involving impersonation of banks, public institutions and prominent local figures. Its training also covered regional fraud patterns intended to make fraudulent exchanges appear credible to Singaporean users.
The system operates through two stages to limit the processing demands placed on smartphones. A lightweight classifier initially examines incoming messages for possible threats, consuming relatively little computing power. Messages considered suspicious then undergo a more detailed assessment using the fine tuned Gemma model.
During the second stage, the AI examines a broader section of the conversation to assess whether the exchange displays patterns of manipulation. This contextual approach is intended to distinguish ordinary conversations from attempts to build trust before requesting payments, passwords or other confidential details.
Where the model identifies potentially fraudulent behaviour, it generates a warning on the handset, encouraging the recipient to pause and exercise caution. The alert is designed to intervene before a person acts on a scammer’s instructions, although the organisations have not disclosed measured reductions in fraud losses from using the system.
A central feature is that the analysis takes place locally. Conversations do not need to be uploaded for cloud based examination, an important distinction for messaging services that use end to end encryption to prevent intermediaries from reading message contents.
Mark Chen, director of trust and safety at GovTech Singapore, said the collaboration sought to identify targeted local scams in real time while keeping personal communications private. The partners present local processing as a way of reconciling fraud prevention with confidentiality.
The privacy safeguards also extend to development. Google DeepMind said the model was fine tuned inside a government managed Google Cloud environment described as a sovereign sandbox, where training material remained isolated.
According to the company, private citizen conversations and personally identifiable information were excluded from training. Instead, developers used anonymised datasets derived from scam reports and intelligence feeds. Testing included synthetic conversations and local material covering different scam categories and changing impersonation methods.
Those measures describe the development process and proposed operation, rather than an independent assessment of the finished product. The partners have not released detailed public benchmarks showing false positive rates, missed scams, battery consumption or performance across different smartphone models.
Singapore has strong reasons to pursue additional protections. Google DeepMind said scams account for more than half of crimes reported in the country, while cumulative losses since 2019 exceed S$4 billion. It estimated that social engineering schemes account for as much as 80 per cent of scam losses.
Such schemes often depend on prolonged interaction, with perpetrators establishing apparent relationships or authority before introducing requests for money. Detecting suspicious intent across multiple messages therefore presents a different technical challenge from blocking known fraudulent websites or telephone numbers.
The announcement comes alongside separate Android security enhancements unveiled by Google, including optional cloud assisted analysis in Google Messages. That service uses a protected computing environment, whereas the Singapore collaboration specifically centres on keeping conversation analysis on the handset.
The distinction is significant because the two approaches involve different arrangements for processing private communications. The Singapore model is designed to assess message content without transmitting conversations externally, although its practical privacy protections will also depend on how individual applications implement the technology.
Google DeepMind said its model was optimised for smartphones commonly used in Singapore. It has not specified the minimum hardware requirements, participating mobile manufacturers or whether the technology will eventually support additional languages and markets.
When Abbie Dibbert and Francoise Albrando took over the day-to-day running of Stay® Property Management, the region was in the middle of a conflict. They chose to keep building. “We saw an opportunity to grow, to expand, and to believe even harder in the UAE and its ability to rise within the chaos,” the directors explain. Over the six months following their March takeover, the live portfolio […]
Google has made its new frontier artificial intelligence model, Gemini 4 Argon, available through Google AI Studio, expanding developers’ access to advanced reasoning, software engineering and autonomous task execution capabilities. The global rollout, announced on October 13, extends access to developers and enterprises through the Gemini API, with the company positioning Argon as its most capable model for complex professional and scientific workloads. Google said the model […]
Meta has begun notifying approximately 1,500 people after attackers accessed its GitHub repositories using stolen employee credentials, obtaining software code and personal information belonging largely to employees in the United States. The company said notifications began on Tuesday, 13 October, several weeks after its security team discovered the intrusion. An outside group had used credentials stolen from an employee to enter repositories containing business information associated with […]
Meta chief executive Mark Zuckerberg has acknowledged that the company contemplated committing as much as $200 billion to artificial intelligence infrastructure before knowing whether its next generation of models would succeed. Speaking on the Cheeky Pint podcast released on Monday, Zuckerberg described discussions late last year with chief AI officer Alexandr Wang about the enormous financial commitment required to build computing capacity ahead of demonstrated technical results. […]
Meta Platforms has launched its Muse Wonder artificial intelligence models, claiming performance comparable with leading systems from OpenAI and Anthropic, marking a significant advance in the company’s efforts to compete at the technological frontier. The company introduced Muse Wonder and Muse Wonder Full on October 12, describing them as its most capable models, designed for complex professional assignments, software development and autonomous digital assistants. Chief executive Mark […]
