
Singapore faces serious cyberattack on critical infrastructure
'Even as we speak, UNC3886 is attacking our critical infrastructure right now,' Shanmugam stated, emphasizing the severity of the advanced persistent threat (APT). The group's operations risk compromising national security by disrupting essential services, including power grids, healthcare networks, and transportation systems.
The minister highlighted the potential cascading effects of a successful breach, noting, 'Our banks, airports, and industries would not be able to operate. Our economy can be substantially affected.' Between 2021 and 2024, suspected APT incidents in Singapore surged over fourfold, underscoring escalating cyber warfare risks.
A 2018 cyber intrusion into a public healthcare cluster, which exposed medication records of 160,000 patients—including former Prime Minister Lee Hsien Loong—was cited as a precedent. Satnam Narang of Tenable warned that defending against such stealthy, well-resourced adversaries is 'increasingly demanding' as digital infrastructure expands.
The Cyber Security Agency (CSA) is actively countering the threat, though Shanmugam withheld details on mitigation strategies. Industry analysts stress that UNC3886's tactics align with long-term espionage campaigns, often targeting governmental and industrial data. - AFP

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles


The Star
4 hours ago
- The Star
Is AI the future of web browsing?
When was the last time you thought about your web browser? If you don't remember, no one will blame you. Web browsers have remained fundamentally unchanged for decades: You open an app, such as Chrome, Safari or Firefox, and type a website into the address bar. Many of us settled on one and fell into what I call 'browser inertia,' never bothering to see if there's anything better. Yet a web browser is important because so much of what we do on computers takes place inside one, including word processing, chatting on Slack and managing calendars and email. That's why I felt excited when I recently tried Dia, a new kind of web browser from the Browser Co. of New York, a startup. The app is powered by generative artificial intelligence, the technology driving popular chatbots like ChatGPT and Google's Gemini, to answer our questions. Dia illuminates how a web browser can do much more than load websites – and even help us learn and save time. I tested Dia for a week and found myself browsing the web in new ways. In seconds, the browser provided a written recap of a 20-minute video without my watching its entirety. While scanning a breaking news article, the browser generated a list of other relevant articles for a deeper understanding. I even wrote to the browser's built-in chatbot for help proofreading a paragraph of text. Dia is on the cusp of an emerging era of AI-powered internet navigators that could persuade people to try something new. This week, Perplexity, a startup that makes a search engine, announced an AI web browser called Comet, and some news outlets have reported that OpenAI, the company behind ChatGPT, also plans to release a browser this year. OpenAI declined to comment. (The New York Times has sued OpenAI and its partner, Microsoft, claiming copyright infringement of news content related to AI systems. The two companies have denied the suit's claims.) Tech behemoths like Google and Apple have added lightweight AI features into their existing browsers, Chrome and Safari, including tools for proofreading text and automatically summarizing articles. Dia, which has not yet been publicly released, is available as a free app for Mac computers on an invitation-only basis. What does this all mean for the future of the web? Here's what you need to know. What is an AI browser, and what does it do? Like other web browsers, Dia is an app you open to load webpages. What's unique is the way the browser seamlessly integrates an AI chatbot to help – without leaving the webpage. Hitting a shortcut (command+E) in Dia opens a small window that runs parallel to the webpage. Here, you can type questions related to the content you are reading or the video you are watching, and a chatbot will respond. For example: – While writing this column on the Google Docs website, I asked the chatbot if I used 'on the cusp' correctly, and it confirmed that I did. – While reading a news article about the Texas floods, I asked the browser's chatbot to tell me more about how the crisis unfolded. The bot generated a summary about the history of Texas' public safety infrastructure and included a list of relevant articles. – While watching a 22-minute YouTube video about car jump starters, I asked the chatbot to tell me which tools were best. Dia immediately pulled from the video's transcript to produce a summary of the top contenders, sparing me the need to watch the entire thing. In contrast, chatbots like ChatGPT, Gemini and Claude require opening a separate tab or app and pasting in content for the chatbot to evaluate and answer questions, a process that has always busted my workflow. How does it work? AI chatbots like ChatGPT, Gemini and Claude generate responses using large language models, systems that use complex statistics to guess which words belong together. Each chatbot's model has its strengths and weaknesses. The Browser Co. of New York said it had teamed up with multiple companies to use their AI models, including the ones behind Gemini, ChatGPT and Claude. When users type a question, the Dia browser analyzes it and pulls answers from whichever AI model is best suited for answering. For instance, Anthropic's AI model, Claude Sonnet, specializes in computer programming. So if you have questions about something you are coding, the browser will pull an answer from that model. If you have questions about writing, the Dia browser may generate an answer with the model that OpenAI uses for ChatGPT, which is well known for handling language. What I appreciate about this design is that you, the user, don't need to know or think about which chatbot to use. That makes generative AI more accessible to the mainstream. 'You should just be able to say, 'Hey, I'm looking at this thing, I've got a question about it,'' said Josh Miller, the CEO of the Browser Co., which was founded in 2020 and has raised over $100 million. 'We should be able to answer it for you and do work on your behalf.' But aren't there imperfections? While Dia proved helpful in most of my tests, it was, like all generative AI tools, sometimes incorrect. While I was browsing Wirecutter, a New York Times publication that reviews products, I asked the chatbot if there were any deals on the site for water filters. The chatbot said no, even as I read about a water filtration system that was on sale. Miller said that because the browser drew answers from various AI models, its responses were subject to the same mistakes as their respective chatbots. Those occasionally get facts wrong and even make things up, a phenomenon known as 'hallucination.' More often than not, however, I found Dia to be more accurate and helpful than a stand-alone chatbot. Still, I double-checked answers by clicking on any links Dia's bot was citing, like the articles about the recent floods in Texas. What about privacy? Asking AI to help with a webpage you're looking at means that data may be shared with whatever AI model is being used to answer the question, which raises privacy concerns. The Browser Co. said that only the necessary data related to your requests was shared with its partners providing AI models, and that those partners were under contract to dispose of your data. Privacy experts have long warned not to share any sensitive information, like a document containing trade secrets, with an AI chatbot since a rogue employee could gain access to the data. So I recommend asking Dia's chatbot for help only with innocuous browsing activities like parsing a YouTube video. But when browsing something you wouldn't want others to know about, like a health condition, refrain from using the AI. This exchange – potentially giving up some privacy to get help from AI – may be the new social contract going forward. How much will this cost? Dia is free, but AI models have generally been very expensive for companies to operate. Consumers who rely on Dia's AI browser will eventually have to pay. Miller said that in the coming weeks, Dia would introduce subscriptions costing US$5 a month to hundreds of dollars a month, depending on how frequently a user prods its AI bot with questions. The browser will remain free for those who use the AI tool only a few times a week. So whether an AI browser will be your next web browser will depend largely on how much you want to use, and pay, for these services. So far, only 3% of the people who use AI every day are paid users, according to a survey by Menlo Ventures, a venture capital firm. That number could grow, of course, if generative AI becomes a more useful tool that we naturally use in everyday life. I suspect the humble web browser will open that path forward. – ©2025 The New York Times Company This article originally appeared in The New York Times.


The Sun
18 hours ago
- The Sun
Plan to transform Kuala Lumpur into Malaysia's flagship AI-driven smart city
KUALA LUMPUR: Malaysia is setting the stage for an artificial intelligence-driven future, beginning with Kuala Lumpur as the flagship city for smart urban governance and seamlessly integrated digital infrastructure. As the nation's capital transforms into a model of intelligent city planning, it signals the country's broader commitment to harnessing AI for sustainable, efficient, and citizen-centric development. Digital Minister Gobind Singh Deo unveiled ambitious plans to transform the capital into a model AI city during a visit to the Kuala Lumpur Command and Control Centre (KLCCC) today, laying the groundwork for a broader 'Smart Nation' strategy under Malaysia Digital 2030. 'It is clear that the focus is on how we can build a command centre that enables us to deploy smart solutions,' Gobind said during a press conference at the KLCCC. 'What is important now is to bring together the existing strengths of each agency and look toward building a National Smart Nation Centre.' The KLCCC, operated by Kuala Lumpur City Hall, serves as the city's nerve centre for smart traffic management, emergency response and urban planning. With AI-driven systems already in place, including real-time traffic signal adjustments, citywide surveillance analytics, and a developing 'digital twin' of Kuala Lumpur, the centre marks a pivotal milestone in Malaysia's digital transformation. The digital twin, a virtual replica of the city, allows for real-time simulation and analysis of urban dynamics. This supports more effective planning, emergency response and resource optimisation. It is one of several AI-enabled features integrated into Kuala Lumpur's urban management framework. A key enabler of the AI-driven city model is the country's data infrastructure. Gobind said the ministry is moving forward with a 'cloud-first' policy and is working to ensure that data across ministries is digitised, centralised and AI-ready. 'We need to ensure that the data we have is in a usable form for AI,' he said. 'That means digitising not just digital data, but also analogue, visual and audio inputs. Once digitised, we need compute power and secure cloud storage to turn data into meaningful solutions.' He pointed to Malaysia's expanding data centre landscape, including hyperscale investments by Microsoft, Google and Alibaba, as crucial infrastructure to support this transition. 'We have a cloud-first policy because that is how we will see new technologies work in AI,' he added. 'When we talk about data, we also talk about storage, security and privacy.' The minister also highlighted the role of the Data Centre Task Force, which aims to streamline infrastructure planning and resource allocation in line with AI and cloud computing needs. During his walkthrough of KLCCC, Gobind stressed the importance of breaking down data silos across ministries to deliver seamless services. 'Each ministry holds data that is useful. The Health Ministry has new systems to reduce waiting times by letting patients pre-book appointments, while the Transport Ministry has real-time data on buses and trains,' he said. 'Imagine combining that data to help someone get to their doctor on time, using the best available route. That's how AI adds value.' Gobind said integration is key to reducing inefficiencies, cutting costs and improving public service delivery. 'We must make sure that the data we have is integrated. There is no need to recreate the wheel. This allows us to move faster and optimise what we already have.' Kuala Lumpur will host the Smart City Expo Kuala Lumpur 2025 from Sept 17 to 19, with co-hosts Malaysia Digital Economy Corporation and Digital Nasional Bhd. The event, a satellite of the global Smart City Expo World Congress held annually in Barcelona, will position KL as a regional hub for smart city innovation. 'We are preparing now. I've been visiting other countries to understand their strengths and to invite them to Kuala Lumpur,' Gobind said. 'We want KL to be the showcase for this region.' The event is expected to attract AI solution providers, tech investors, and city planners, making KL a bridge for Asean digital collaboration. Gobind concluded by reinforcing the government's long-term vision under Malaysia Digital 2030, which aims to evolve from isolated smart city projects into an integrated smart nation ecosystem. 'We must always be ready for the technology of the next five years,' he said. 'If we move ahead without preparation, we risk falling behind. That's why the Prime Minister's digital transformation vision is focused on readiness, trust and secure systems.'


Malaysia Sun
18 hours ago
- Malaysia Sun
U.S. reviews military cloud deals after China-linked tech report
SAN FRANCISCO, California: Microsoft announced July 18 it will no longer allow engineers based in China to provide technical assistance for U.S. military systems, following scrutiny from a U.S. senator and a newly ordered review by Defense Secretary Pete Hegseth into Pentagon cloud contracts. The move comes after investigative outlet ProPublica reported that Chinese engineers were supporting U.S. military cloud computing systems under the supervision of U.S. "digital escorts"—subcontractors with security clearances but often lacking the expertise to evaluate cybersecurity risks. The report raised concerns about potential vulnerabilities and prompted swift action from lawmakers and the Pentagon. Microsoft, one of the largest technology contractors to the U.S. government, confirmed that the arrangement had been disclosed to federal authorities during its contract authorization process. Still, in response to the backlash, the company has now pledged to overhaul its procedures. "In response to concerns raised earlier this week, we've changed how we support U.S. government customers to ensure that no China-based engineering teams are providing technical assistance," Microsoft spokesperson Frank Shaw said on X (formerly Twitter). Earlier in the day, Senator Tom Cotton, who chairs the Senate Intelligence Committee and serves on the Armed Services Committee, sent a letter to Defense Secretary Hegseth demanding clarity about contractors using personnel from China. He also requested details about how digital escorts are trained to detect security breaches. "The U.S. government recognizes that China's cyber capabilities pose one of the most aggressive and dangerous threats to the United States," Cotton wrote. He cited prior intrusions into critical infrastructure and telecom networks as justification for increased vigilance. "The U.S. military must guard against all potential threats within its supply chain, including those from subcontractors." Defense Secretary Hegseth responded swiftly, ordering a full two-week review of all Defense Department cloud service contracts to determine if any others involve China-based personnel. "I'm announcing that China will no longer have any involvement whatsoever in our cloud services, effective immediately," Hegseth said in a video posted online. "We will continue to monitor and counter all threats to our military infrastructure and online networks." The Pentagon has awarded billions in cloud computing contracts to major tech companies as part of its modernization strategy. Cybersecurity has been a growing concern after a series of high-profile hacks—some traced back to state-sponsored actors in China and Russia—targeted U.S. government and corporate systems, including Microsoft's own networks. While Microsoft maintains that proper disclosure protocols were followed, the company's use of Chinese engineers—even under supervision—has reignited debate over foreign access to sensitive military systems. The review launched by Hegseth will assess whether additional safeguards are necessary and whether similar practices have occurred with other vendors.