Singapore AI News & Daily Briefing

Singapore is pushing responsible AI adoption in schools, integrating the technology thoughtfully into education so students learn to use it safely and effectively. The government is balancing innovation with safeguards, ensuring AI tools enhance learning without creating new risks for young users.

Singapore has updated its Retail Industry Digital Plan to help small and medium businesses adopt technology faster. This means local shops and retailers could soon get easier access to AI tools, digital payment systems, and smart inventory management — potentially lowering costs and improving customer service.

Nvidia CEO Jensen Huang says fears that AI will destroy jobs are 'complete nonsense'. He argues AI makes software engineers more productive, so companies hire more of them. As AI becomes profitable, demand for developers grows. He also unveiled a new AI superchip for PCs, built with Microsoft and MediaTek, that can run powerful AI models on your laptop without internet.

SK Hynix, the world's second-largest memory chipmaker, plans to double its production capacity over the next five years to tackle a persistent global shortage of memory chips needed for AI. Chairman Chey Tae-won says the shortfall could last until 2030, and the company is prepared to spend whatever it takes to expand. SK Hynix controls about 60 percent of the high-bandwidth memory (HBM) market alongside Samsung and Micron, and has filed to list American depositary receipts in New York to give US investors easier access to the AI memory boom.

OpenAI is pushing for a global youth AI safety institute to be established, presenting principles at the G7 Leaders' Summit in France. The framework covers age-appropriate protections, parental controls, limits on data collection from minors, and safeguards against harmful content. OpenAI highlights Singapore as a partner country already working with them on AI education initiatives in schools.

A critical vulnerability called BadHost in the widely-used Starlette package lets hackers break into servers running AI agents. This can expose sensitive data and credentials. Millions of AI tools are affected.

As AI systems become more capable, a lot of resources and effort are being put toward measuring their abilities. Researchers look at technical evaluation metrics, subject AIs to reasoning tests, track their throughput, and much more. But there’s one key metric that often gets overlooked, and it’s arguably the most important of all: What is AI doing to humans? Imran Khan leads psychosocial evaluation of AI at the nonprofit Center for Humane Technology . In a recent essay published on the organization’s Substack, Khan points out that we’re deploying AI tools capable of reshaping our cognition, relationships, and behavior, but with little systematic effort to measure the downstream impacts they’re having on us. The push to look more closely at AI’s psychosocial effects is similar to debates t

Travelers, a major US insurance company, has rolled out an AI-powered voice assistant nationwide to handle auto insurance claims. The system talks customers through filing claims after car accidents, answering questions and submitting paperwork without wait times. Over 85-90% of users now complete their claims through the AI, which handles surge events that can generate 100,000+ claims in days.

OpenAI has expanded Codex beyond coders to help non-technical workers like marketers, analysts, and bankers. New features include role-specific plugins that connect to tools like Salesforce, Figma, and Snowflake, plus the ability to create shareable websites and dashboards directly from conversations. Over 5 million people now use Codex weekly, with non-developers growing 3x faster than developers.

AI can now handle basic admin tasks like invoicing, social media planning, and summarizing meetings for small businesses. This means even solo entrepreneurs can automate tedious paperwork without hiring extra staff.

This sponsored article is brought to you by Master Bond . Outgassing is the release of volatile substances from a cured adhesive over time. These released materials, which may include residual solvents, unreacted monomers, or other chemical species, can deposit on nearby surfaces, causing contamination that interferes with sensitive components. What Is Outgassing and How Is It Measured? The industry standard for measuring outgassing is ASTM E595, developed by NASA . This test exposes a cured sample to 125 °C at high vacuum (10⁻⁵ to 10⁻⁶ torr) for 24 hours, measuring Total Mass Loss (TML) and Collected Volatile Condensable Materials (CVCM). To meet NASA low outgassing requirements, materials must exhibit less than 1 percent TML and less than 0.1 percent CVCM. Optical assemblies need contami

A new report from OpenAI explores how Codex is reshaping knowledge work across industries. The research examines how AI-powered research, data analysis, workflow automation, and content creation tools are changing productivity for non-technical workers. Over 5 million people now use Codex weekly, with non-developer adoption growing faster than developer adoption.

“Not in my backyard” is the rallying cry of citizens everywhere resisting projects proposed for their locality. Whether it’s affordable housing, a waste treatment plant, or a new data center, they may recognize the benefit of the activity. They just don’t want it near them. And the roots of that resistance differ from place to place. When it comes to the ongoing transition from fossil fuels to renewables, companies and policymakers need to know where, exactly, people are coming from. The Italian island of Sardinia is a textbook example. As IEEE Spectrum ’s power and energy editor Emily Waltz discovered when she traveled there last October, Sardinian opposition to wind and solar projects runs deep. It spurred a quarter of the voting population to queue up in public squares in 2024 to sign a

OpenAI has publicly disclosed that it does not have a Political Action Committee, has not donated to any super PACs, and does not fund political candidates. The company clarifies that no outside group — including Leading the Future, backed by co-founder Greg Brockman — speaks for OpenAI. This transparency push comes as AI policy becomes a hot-button political issue, with OpenAI stating it prefers to advocate directly and openly rather than through shadowy lobbying groups.

Red Hat's official NPM account was compromised and used to push malicious packages containing a worm called Shai-Hulud that steals sensitive credentials including GitHub secrets, npm tokens, and cloud service passwords. The malware spreads by republishing backdoored versions to other accounts the infected system can access. Anyone who downloaded these packages in the past 36 hours should assume their systems and credentials are compromised.

America's top cybersecurity agency CISA has had a major security failure: a contractor accidentally left sensitive credentials—including passwords, SSH keys, and tokens—exposed in a public GitHub repository since November 2025. A researcher was able to use these credentials to access US government cloud systems at high privilege levels. This follows another CISA security mishap earlier this year when the agency's acting director uploaded classified documents to ChatGPT. The lesson here is simple: even the people responsible for protecting government systems can make basic security mistakes that put everyone at risk.

Apple finally unveiled a major overhaul of Siri at its annual developer conference, adding more conversational AI, screen awareness, and web search capabilities. The new "Siri AI" can find information mentioned casually in messages and answer questions about what appears on your screen. Apple is using Google's Gemini technology for some features and Nvidia chips for cloud processing, while promising that most data stays private on your device. The update is unavailable in the EU and China for now due to regulatory hurdles.

OpenAI has quietly filed for a US IPO, following Anthropic's move toward public markets. The filing signals that the world's most high-profile AI company is preparing to raise capital from public shareholders, shifting from its non-profit startup origins. Details on valuation and timeline remain confidential.

Apple unveiled a rebuilt Siri with new AI capabilities at its WWDC conference, but investors were unimpressed — the stock fell 1.9% during the keynote. The new Siri can tap into your personal data, control multiple apps, and hold natural conversations, but it will launch as a beta in English only this autumn. Advanced features won't work on phones older than the iPhone 14, and won't be available in China or the EU at launch. Apple also agreed to pay US$250 million to settle claims it exaggerated earlier AI features.

Singtel has partnered with Digital Industry Singapore (DISG) for a multi-year AI transformation program. The initiative will touch operations, digital infrastructure, customer platforms, and workforce training across Singtel's businesses.

Apple has rebuilt Siri from the ground up with nine major new capabilities coming in iOS 27 this autumn. Key upgrades include a dedicated Siri app with conversation history, cross-app actions (like finding photos and drafting emails in one go), a camera-based Visual Intelligence mode that reads receipts and identifies food, system-wide writing assistance, and deeper integration with smart glasses and Vision Pro. Developers can now hook their apps into Siri — Apple cited Line (widely used in Southeast Asia) as an early partner. The new Siri requires iPhone 15 Pro or later to use.

Intel shares have jumped 65% this year after winning deals with Google, which ordered 3 million TPUs. Google will use Intel's Xeon 6 processors for AI work and join Terafab, a Texas project building custom chips for Elon Musk's companies. Intel also bought back full control of its Ireland chip plant for $14.2 billion. The deals show big tech companies spreading AI spending beyond Nvidia to build balanced systems.

Elon Musk says SpaceX's AI-powered satellites will mostly use existing technology when the company goes public. The comment comes ahead of a potential SpaceX IPO, where Musk is positioning AI capabilities as part of the satellite business pitch to public market investors.

US lawmakers are moving closer to creating clear rules for stablecoins — cryptocurrencies designed to stay equal to the US dollar. Major retailers including Walmart and Amazon are exploring stablecoins to cut high credit card fees and speed up payments. Visa and Mastercard are adapting by offering stablecoin services. While $27.6 trillion in stablecoin transfers happened in 2024 — more than Visa and Mastercard combined — consumer adoption still faces hurdles like complicated wallet setup and lack of reward programs.

Maybank says AI is reshaping Southeast Asia's internet sector, moving beyond cost savings to build platform capabilities around payments, logistics and consumer trust. Singapore-listed Sea Ltd is flagged as a key beneficiary of this shift. The bank sees Sea well-positioned as AI transforms how digital services compete in the region.

Singapore's sovereign wealth fund Temasek is leading a $300 million funding round for UK AI company PhysicsX, which uses AI to speed up engineering simulations. The company creates AI models that can predict physical behavior in seconds instead of hours or days, helping aerospace, semiconductor, and automotive companies design better products faster. PhysicsX has doubled its revenue and tripled its bookings over the past year.

Singapore-based semiconductor firm Silicon Box has raised S$100 million ($78 million) in debt financing from investors including Ares Management. The company specializes in advanced chip packaging technology, which helps pack more computing power into smaller spaces — critical for AI chips and data centers.

Perplexity AI has made its Comet browser available to everyone for free, after initially limiting it to paid subscribers. The browser can search the web, help with research, and handle multiple tasks at once. CEO Aravind Srinivas claims it could boost productivity enough to add trillions to global GDP, though economists say the job impact so far is limited. Major tech firms like Meta, Amazon, Microsoft, and Google are all racing to build similar AI-powered browsers.

SpaceX has filed for an IPO that could value the company at up to $2 trillion, which would make Elon Musk the world's first trillionaire. The filing shows Musk holds about 5.1 billion shares plus options. He can earn billions more in stock awards if SpaceX hits valuation targets tied to building permanent Mars colonies and non-Earth data centers capable of delivering 100 terawatts of computing power yearly — essentially tying AI computing expansion to space ambitions.

Apple is spending more on AI than it has in three decades, allocating 10.3% of its revenue to R&D. The company is working to catch up with Google and Microsoft on AI capabilities. Siri and Apple Intelligence upgrades are coming later in 2026, and Apple is partnering with Google to use its Gemini AI model for Siri — a deal reportedly worth around $1 billion yearly. Apple is betting that AI models will eventually become interchangeable commodities, making user experience and privacy the key differentiators.

Asian tech stocks fell sharply as investors started questioning whether AI companies can actually deliver on their promises. South Korea's Kospi saw its steepest drop in months, triggering automatic trading halts. Samsung and SK Hynix both took big hits. Even Singapore's Straits Times Index slipped 1.7% with chip stocks among the worst performers. Analysts say concerns about AI spending momentum and high valuations are spooking markets, combined with geopolitical tensions in the Middle East.

Despite recent market wobbles, analysts say Malaysia's tech sector is entering a growth phase driven by actual earnings rather than just hype. Improving order books and sustained global AI infrastructure spending are supporting the outlook, though experts warn of increased volatility ahead.

Nvidia and Microsoft are developing specialized computer hardware designed specifically for AI agents — software programs that autonomously perform tasks. Current computer chips weren't built for this type of work, so both companies are creating custom solutions to make AI agents faster and more efficient.

Singapore is seeing a sharp rise in white-collar layoffs, with retrenchments jumping from 6,440 in 2022 to 14,400 in 2025. More affected professionals — especially in finance, IT, and professional services — are now turning to unions for help, though many still don't know unions can represent them. AI and offshoring are cited as key drivers behind the restructuring, with some companies relocating roles out of Singapore even when local workers are experienced. NTUC handled over 3,900 retrenchment cases for PMEs last year, up 5% from 2024. Labour economists are calling for stronger safeguards, including better unemployment benefits and retrenchment insurance schemes.

SK Telecom is partnering with Korean hardware startup Panmnesia to redesign AI data centers using a technology called CXL, which lets servers share memory and processing power more flexibly across racks. The goal is to cut waste by letting CPUs, GPUs, and memory be dynamically allocated wherever needed, avoiding expensive idle hardware. They also want to replace standard network connections between GPUs with faster CXL links to boost AI processing speed without adding more chips.

Samsung spent a record $25.7 billion on research and development in 2025, up 8% from the previous year, to accelerate AI chip development including high-bandwidth memory and next-generation DDR5 chips. The Korean electronics giant is fighting to reclaim market share in AI memory after rival SK Hynix overtook it in annual profits for the first time, partly thanks to SK Hynix's lead in supplying HBM chips to Nvidia. Samsung also invested $35.9 billion in facilities, including new semiconductor plants, as memory costs for data centers surge.

Chinese cloud stocks rallied after a local government in Shenzhen drafted policies supporting the open-source AI agent platform OpenClaw, with shares jumping 9% or more for firms like UCloud and QingCloud. However, China's Ministry of Industry and Information Technology flagged serious security risks from improper configurations in OpenClaw deployments, warning that poorly set-up systems could leak data or get hacked. The buzz shows how AI platforms are reshaping China's cloud market beyond basic pricing wars, but it also highlights the real security headaches that come with rapid AI adoption.

Japanese telecom giant NTT is planning to raise approximately US$864 million through a data center Real Estate Investment Trust (REIT) listing on the Singapore Exchange. The REIT will hold six data center assets across the US, Austria, and Singapore. Singapore's sovereign wealth fund GIC has committed $101 million as a cornerstone investor. The listing, valued at up to US$1.08 billion, represents the largest IPO on SGX since 2021 and signals renewed confidence in Singapore's market following recent reforms to attract listings.

A global study by Tata Communications and Bloomberg Media Studios finds that while 75% of enterprise leaders now treat AI as a board-level priority, only 29% are confident their infrastructure can meet evolving demands. The survey of 501 senior executives at companies with over $500 million in revenue reveals that nearly two-thirds are still operating on legacy infrastructure not designed for AI's data intensity. Key barriers include integration challenges with legacy systems (cited by 27%), talent shortages (45% among firms over $5 billion revenue), and lengthy governance reviews. Companies with fully modernized infrastructure are nearly twice as likely to report meaningful business value from AI.

Apple finally unveiled its revamped Siri AI at WWDC 2026, two years after originally promising these features. The new Siri uses Google's AI models and can understand what's on your screen, help across apps, and handle writing tasks. Industry experts are divided — some say it could threaten standalone AI companies since Apple bundles these features free with devices, while others note Apple still lacks access to third-party app data that would make the AI truly personalized. The company faced a class action lawsuit over its delayed 2024 promises, so the pressure is on to actually deliver.

AI coding startup Cursor, which SpaceX has an option to acquire for $60 billion, is opening its European headquarters in London and plans to hire about 200 staff across the region. The company lets developers write code by describing apps in plain language. Cursor reports around $2.6 billion in annualized revenue and counts British Airways, BP, and Nokia among its clients.

Tech giant NTT DATA has signed a major multi-year deal with Amazon Web Services to help businesses move to the cloud and adopt AI tools. The partnership brings together nearly 11,000 AWS-certified experts from NTT DATA to help companies modernize old computer systems, build industry-specific solutions, and use AI responsibly. Key areas include finance, healthcare, and manufacturing.

Amazon Web Services has committed to investing US$7 billion over 14 years to expand its cloud data center infrastructure in Telangana, India. The new Hyderabad region will support cloud services for startups, enterprises, government agencies, and AI projects across India. AWS already operates in Mumbai and Hyderabad, and this expansion strengthens its India presence.

A new survey shows Singapore employers remain willing to pay more for AI-skilled workers even as overall hiring sentiment softens. Workers with AI literacy, critical thinking, and problem-solving abilities continue to command salary premiums in the local job market, despite broader economic uncertainty.

TSMC has become one of the world's top 10 most valuable companies, with its market value surging 101% year-on-year to US$1.427 trillion. This jump is driven by insatiable global demand for AI chips and semiconductors. TSMC fabricates chips for Apple, Nvidia, and countless other tech giants.

Alibaba Cloud is expanding its AI infrastructure by launching new data centers across eight countries, including Malaysia. This move aims to meet growing demand for cloud computing and AI services in Southeast Asia, intensifying competition among cloud providers in the region.

New York City is now requiring advertisers to clearly label AI-generated 'synthetic performers' in ads. This means if a brand uses a computer-generated influencer or AI-created avatar in their advertising, they must disclose it upfront. The rule aims to protect consumers from being misled about who's actually endorsing a product.

Singapore Post (SingPost) has unveiled a new automated parcel sorting hub featuring high-tech automation systems. The facility represents a significant upgrade to Singapore's logistics infrastructure, likely incorporating AI-driven sorting and robotics to handle growing e-commerce volumes more efficiently.

Singapore is setting up a dedicated test area in Punggol where companies like Grab, DHL, and Certis can try out robots for delivering food, parcels, and patrolling public spaces. IMDA and SIT are working with industry partners including Nvidia to build real-world robotics conditions. The government also updated its AI governance rules for agentic AI systems.

A robot called 'Robbie' (a Stretch 4 model) helps dementia patient Brian Marquis with daily tasks—reminding him to eat, fetching water, reading medication labels. His wife Brenda can now leave the house without worry. This addresses the growing eldercare crisis—by 2030, Americans over 65 will outnumber those under 18 for the first time. The robot frees caregivers from repetitive tasks while keeping patients safe at home.

Anthropic has launched Claude Fable, the first public version of its Mythos-class AI model. The company built in major safeguards to address safety concerns about releasing a more powerful model to the masses. Users can now access capabilities previously limited to select testers. The move puts pressure on competitors like OpenAI and Google to demonstrate their models remain competitive.

Samsung has started shipping samples of its new HBM4E memory chips to major AI chip customers. The 12-layer chip offers 48GB capacity and up to 3.6TB/s bandwidth, running over 20% faster than the previous generation. Next-generation AI processors like Nvidia's Rubin are expected to use this memory type.

Anthropic released Claude Fable 5, a new Mythos-class AI model now available to the public. It's the latest addition to the Claude family competing in the crowded AI assistant market.

Philips North America's CEO says AI is meaningfully improving accuracy for doctors and nurses in clinical settings. The technology helps reduce diagnostic errors and assists healthcare workers in making faster, more informed decisions. This signals a shift from AI hype to real-world medical applications, though questions remain about implementation costs and doctor adoption.

SpaceX reportedly plans to test AI computing capabilities in orbit by the end of next year, according to unnamed sources. Orbital AI computing could process data with lower latency and potentially avoid terrestrial data center constraints. Few details are available yet.

Alibaba Cloud has opened two new data centers in Johor, Malaysia—its largest infrastructure investment in Southeast Asia. The expansion comes with a suite of agentic AI services, essentially AI systems that can autonomously handle tasks, being offered to Malaysian businesses from the second half of this year.

Singapore insurtech Igloo has launched an AI assistant called Igi in Indonesia that guides users through buying travel insurance via a conversational interface. Early results show users who chat with Igi are 44% more likely to complete a purchase compared to browsing the website alone. Indonesia has extremely low insurance penetration at just 2.7%, and Igloo aims to make buying coverage as easy as ordering food delivery. The AI handles the entire sales journey from plan selection to payment, only handing off to human agents when needed.

A new security report reveals that 18% of Singapore organizations have policies banning unauthorized AI tools but have no way to actually detect violations—the highest rate in Asia Pacific. Meanwhile, Singapore leads in other enforcement areas like network monitoring at 67% and AI code scrutiny at 71%. The findings highlight a gap between having security rules on paper and actually enforcing them, especially as malicious software packages and weaponized AI models flood public registries.

Malaysian AI firm Zetrix AI's Philippines joint venture is rolling out digital payment options for multiple government agencies including the Professional Regulation Commission, Philippines Ports Authority, and Bureau of Fisheries. Filipinos can now pay fees for professional licenses, port services, and fisheries permits through over 100,000 channels including e-wallets, bank transfers, and QR payments. The goal is eliminating trips to government offices for routine payments.

A major global survey by Philips finds that AI is genuinely saving doctors and nurses time — clinicians report saving an average of 132 hours per year — and helping them see more patients. AI is being used to warn about dangerous drug interactions, suggest diagnoses, and analyse scans. But 70% of healthcare professionals say their organisations are not providing adequate AI training, so many are using personal AI tools instead of workplace ones. Nearly all clinicians — 90% — insist a human must remain in the loop for AI decisions.

An AI cloud executive argues that cheaper AI doesn't shrink the market—it expands it. As computing costs drop, more business tasks become economically viable to automate, driving higher demand for compute rather than less. The release of open-source models like DeepSeek actually boosted sales for some providers even as investor panic caused stock drops. The industry is shifting from selling raw hardware to offering managed AI services that let businesses use AI without building complex systems.

[Nokia](https://www.nokia.com/) and [Indosat Ooredoo Hutchison](https://ioh.co.id/) will jointly modernize Indosat’s nationwide mobile network in Indonesia through the deployment of advanced 5G Radio Access Network technologies, with [NVIDIA](https://www.nvidia.com/en-us/) developing AI-integrated network capabilities. In a statement on Tuesday, Nokia said it would deploy low-band 5G across the full network, with mid-band 5G covering approximately 80 percent of the network over the next three and a half years. Nokia will deploy its Habrok and Pandion radio families, Levante basebands, centralized RAN, and advanced network management and automation platforms. The three companies are also advancing AI-RAN technology, which combines connectivity and artificial intelligence within the network

Indonesian telecom Indosat has partnered with Nokia and Nvidia to open Asia's first AI-RAN (AI Radio Access Network) research center in Surabaya. The facility will develop AI-powered wireless networks, work on energy efficiency, and create an 'AI Grid' connecting AI hubs across Indonesia's 5G network. It aims to build local digital talent while advancing the country's digital transformation goals.

China is preparing a massive $295 billion fund to build out AI infrastructure nationwide, per Bloomberg reports. This is one of the largest government-led AI investment plans globally and could significantly shift the global AI race, potentially putting pressure on other nations to accelerate their own AI strategies.

Japanese tech giant NTT DATA is partnering with Google Cloud to help businesses move AI from pilot projects to full-scale deployment. The plan: certify 5,000 AI experts and build 500 ready-made AI agents for industries like banking, manufacturing, and retail. A survey found 99% of enterprises say AI is driving more cloud spending, but many say their current cloud setup can't keep up with AI demands.

Huawei announced its Ascend supernode, built on Ascend 950 AI chips, will fully support DeepSeek V4. The Chinese AI startup released its largest model yet, claiming it can match OpenAI and Google DeepMind's best closed models. DeepSeek V4 has about 1 trillion parameters and supports a 1 million token context window at lower cost.

Apple failed to get its AI tool to comply with EU regulations, according to the European Commission. This means Apple could face restrictions or penalties in Europe, and its AI features may work differently—or not at all—for EU users compared to other regions. For everyday consumers, this could mean Apple Intelligence features on iPhones may be limited or unavailable in Europe, potentially creating a two-tier experience for iPhone users worldwide.

The first bank account that Isabel Bautista personally signed up for was with GoTyme. “When I saw they were offering interest rates worth 5%, and that cash-ins were free of charge, signing up was as easy as ABC,” she tells _Tech in Asia_. She didn’t stop there. As of 2026, she has three accounts with three different digital banks, all of which are holding upwards of six-figures of her savings. Ever since the country’s central bank – Bangko Sentral ng Pilipinas (BSP) – issued digital bank licenses in late 2020, Filipinos have been spoiled for savings account choices. Today, the six digital bank license holders in the Philippines each offer savings account interest rates that are higher than 3%. Traditional banks in the country offer an average of 1.8%. Image credit: Timmy Loen Though Filipi

Researchers have created an AI-powered ultrasound wristband that can read muscle and nerve signals to let users control robotic hands and virtual reality environments in real time. The device uses sound waves to capture detailed muscle activity, offering more precise control than current wearable sensors. This could eventually lead to smoother prosthetic limbs, better AR/VR experiences, and new hands-free ways to interact with computers.

The European Union has ordered Meta to let rival AI chatbot makers access WhatsApp while regulators investigate whether the company broke competition law. Meta blocked competitors from offering their AI assistants on the platform, raising antitrust concerns. The EU says AI assistants are becoming a major way Europeans will use AI, so blocking competition now could cause lasting harm. Meta plans to appeal the order, which stays in place until 2029 or until the investigation ends.

Anthropic told users they could opt out of having their conversations used to train future AI models, but its updated privacy policy apparently contains a carve-out that contradicts that promise. Users on Free, Pro, and Max plans were asked to choose whether their chats could train Claude—but the new policy may allow Anthropic to use that data anyway. The discrepancy raises questions about whether AI companies are being honest about how they handle your personal conversations.

Healthcare organizations are facing record cyberattacks, with breaches costing an average of $7.42 million per incident. AI is making things worse—hackers use it to craft convincing phishing attacks and find network weaknesses faster. Meanwhile, medical devices like infusion pumps are often poorly secured and connected wirelessly, creating easy entry points. The FBI recorded 642 healthcare cyber incidents in 2025 alone. Experts say hospitals must treat cybersecurity as a patient safety issue, not just an IT problem.

Anthropic released Claude Fable 5, a new Mythos-class AI model now available to the public. It's the latest addition to the Claude family competing in the crowded AI assistant market.

Singapore Post (SingPost) has unveiled a new automated parcel sorting hub featuring high-tech automation systems. The facility represents a significant upgrade to Singapore's logistics infrastructure, likely incorporating AI-driven sorting and robotics to handle growing e-commerce volumes more efficiently.

Anthropic has launched Claude Fable, the first public version of its Mythos-class AI model. The company built in major safeguards to address safety concerns about releasing a more powerful model to the masses. Users can now access capabilities previously limited to select testers. The move puts pressure on competitors like OpenAI and Google to demonstrate their models remain competitive.

Philips North America's CEO says AI is meaningfully improving accuracy for doctors and nurses in clinical settings. The technology helps reduce diagnostic errors and assists healthcare workers in making faster, more informed decisions. This signals a shift from AI hype to real-world medical applications, though questions remain about implementation costs and doctor adoption.

A robot called 'Robbie' (a Stretch 4 model) helps dementia patient Brian Marquis with daily tasks—reminding him to eat, fetching water, reading medication labels. His wife Brenda can now leave the house without worry. This addresses the growing eldercare crisis—by 2030, Americans over 65 will outnumber those under 18 for the first time. The robot frees caregivers from repetitive tasks while keeping patients safe at home.

SpaceX reportedly plans to test AI computing capabilities in orbit by the end of next year, according to unnamed sources. Orbital AI computing could process data with lower latency and potentially avoid terrestrial data center constraints. Few details are available yet.