Top of the Day
The case for expanding digital public infrastructure
(Rowan Wilkinson, Alex Krasodomski – Chatham House . 28 October 2025) The modernization of societies through digital technologies is a critical strategic priority that can improve public services and encourage private sector innovation. The term digital public infrastructure (DPI) refers both to the approach and principles behind open, interoperable digitalization, and to the resulting digital systems and platforms – including those for digital identification, payments and data transfers. The DPI approach focuses on cost-effective solutions, such as using open-source software, open technologies and cross-government interoperability. While some new technologies may raise concerns about privacy and civil liberties, it is possible to address these issues through transparency, democratic oversight and strong legal safeguards. – https://www.chathamhouse.org/2025/10/case-expanding-digital-public-infrastructure
Authoritarian reach and democratic response: A tactical framework to counter and prevent transnational repression
(Marcus Kolga, Sze-Fung Lee, Iria Puyosa, Kenton Thibaut, and Lisandra Novo – Atlantic Council – 27 October 2025) Foreign interference (FI) and transnational repression (TNR) represent a fundamental challenge to the international rules-based order by employing tactics that exist below the threshold of armed conflict while violating national sovereignty. Beyond national borders, authoritarian states have targeted policymakers, elected officials, researchers, journalists, activists, and diaspora communities worldwide to advance their political objectives. These TNR tactics encompass cross-domain operations, including surveillance, cyberattacks, disinformation, legal and judicial harassment, and physical and psychological assault. – https://www.atlanticcouncil.org/in-depth-research-reports/report/authoritarian-reach-and-democratic-response-a-tactical-framework-to-counter-and-prevent-transnational-repression/
DeepSeek Use in PRC Military and Public Security Systems
(Sunny Cheung, Kai-shing Lau – The Jamestown Foundation – 27 October 2025) Military procurement documents show that the People’s Liberation Army (PLA) is adopting homegrown artificial intelligence (AI) systems like DeepSeek to accelerate its shift toward “intelligentized warfare.”. PLA experts describe DeepSeek not as a single product but as an evolving system architecture. They envision integrating this system across the PLA’s entire command, control, communications, computers, intelligence, surveillance, and reconnaissance (C4ISR) chain. Debate is ongoing as to DeepSeek’s utility for PLA purposes. Some powerful institutions back its deployment, while others remain sceptical. Public security and policing are two areas also exploring the use of DeepSeek, especially for enhancing surveillance and data analysis, as well as for assisting with report writing and other processes. The model’s success is framed as both a technological breakthrough and a political achievement in “algorithmic sovereignty.” – https://jamestown.org/program/deepseek-use-in-prc-military-and-public-security-systems/
Nation-State Cyber Ecosystems Weakened by Sanctions, Report Reveals
(Kevin Poireault – Infosecurity Magazine – 28 October 2025) Cyber-related sanctions alone do not typically disrupt cyber malicious activities, but they can “toxify” networks of malicious actors, according to new research. A report, published on October 28 by the Royal United Services Institute (RUSI), builds from the first meeting of the RUSI Cyber Sanctions Taskforce in September. This meeting saw current and former government officials from the UK, the US and the EU, as well as other EU officials, discuss the role of sanctions in countering cyber state threats. – https://www.infosecurity-magazine.com/news/nation-state-cyber-weakened/
NVIDIA, Oracle to build America’s biggest AI supercomputer using 100K Blackwell GPUs
(Interesting Engineering – 28 October 2025) An unprecedented collaboration between NVIDIA and Oracle is set to redefine the frontier of scientific discovery. The companies are joining forces with the U.S. Department of Energy (DOE) to build the nation’s largest AI supercomputer, a system that promises to accelerate innovation across fields from climate science to national security. The Solstice system will feature a record-breaking 100,000 NVIDIA Blackwell GPUs, while a companion machine, Equinox, will pack 10,000 GPUs. – https://interestingengineering.com/energy/nvidia-oracle-doe-ai-supercomputer-nvqlink-palantir
Governance
FDA and patent law create dual hurdles for AI-enabled medical technologies
(DigWatch – 28 October 2025) AI reshapes healthcare by powering more precise and adaptive medical devices and diagnostic systems. Yet, innovators face two significant challenges: navigating the US Food and Drug Administration’s evolving regulatory framework and overcoming legal uncertainty under US patent law. These two systems, although interconnected, serve different goals. The FDA protects patients, while patent law rewards invention. – https://dig.watch/updates/fda-and-patent-law-create-dual-hurdles-for-ai-enabled-medical-technologies – https://www.mondaq.com/unitedstates/patent/1695588/ai-enabled-tech-faces-dual-hurdles-in-fda-regulation-patent-law?email_access=on
Geostrategies
Qualcomm and HUMAIN power Saudi Arabia’s AI transformation
(DigWatch – 28 October 2025) HUMAIN and Qualcomm Technologies have launched a collaboration to deploy advanced AI infrastructure in Saudi Arabia, aiming to position the Kingdom as a global hub for AI. Announced ahead of the Future Investment Initiative conference, the project will deliver the world’s first fully optimised edge-to-cloud AI system, expanding Saudi Arabia’s regional and global inferencing services capabilities. In 2026, HUMAIN plans to deploy 200 megawatts of Qualcomm’s AI200 and AI250 rack solutions to power large-scale AI inference services. – https://dig.watch/updates/qualcomm-and-humain-power-saudi-arabias-ai-transformation – https://www.qualcomm.com/news/releases/2025/10/humain-and-qualcomm-to-deploy-ai-infrastructure-in-saudi-arabia-
Security and Surveillance
Ransomware payments hit record low: only 23% Pay in Q3 2025
(Pierluigi Paganini – Security Affairs – 28 October 2025) Cybersecurity firm Coveware reports that only 23% of ransomware victims paid attackers in Q3 2025, the lowest rate ever recorded. The researchers note this continues a six-year decline in payment rates. After 28% of victims paid in early 2024, rates briefly rose before dropping again, showing that fewer breached companies are paying the ransom. The average ransom payment dropped to $376,941 in Q3 2025 (down 66% from Q2), while the median fell to $140,000 (down 65%). Large firms are refusing to pay, realizing ransoms rarely stop data leaks. Meanwhile, groups like Akira and Qilin target mid-sized firms with smaller, more frequent demands, using a high-volume, low-demand model that’s proving more effective against less resilient victims. “Ransom payment rates across all impact scenarios — encryption, data exfiltration, and other extortion — fell to a historical low of 23% in Q3 2025.” reads the report. “This continuation of the long-term downward trend is something all industry participants should take a moment to reflect on: that cyber extortion’s overall success rate is contracting.” – https://securityaffairs.com/183941/cyber-crime/ransomware-payments-hit-record-low-only-23-pay-in-q3-2025.html
Italian political elite targeted in hacking scandal using stolen state data
(DigWatch – 28 October 2025) Italian authorities have uncovered a vast hacking operation that built detailed dossiers on politicians and business leaders using data siphoned from state databases. Prosecutors say the group, operating under the name Equalize, tried to use the information to manipulate Italy’s political class. – https://dig.watch/updates/italian-political-elite-targeted-in-hacking-scandal-using-stolen-state-data – https://www.politico.eu/article/italy-milan-hackers-carmine-gallo-enrico-pazzali-samuele-calamucci-equalize-mercury-advisors/
Swedish power grid operator confirms it was hit by hacker attack
(Cybernews – 28 October 2025) Svenska kraftnät, Sweden’s primary electricity grid operator, has confirmed that it suffered a data breach after the Russia-linked Everest ransomware gang claimed to have syphoned hundreds of gigabytes of the company’s data. Svenska kraftnät learned about the hacker attack on Saturday, after a security expert notified the company that Everest had posted the company’s data on the gang’s dark web leak site. Cybercartels use these types of underground websites to showcase their latest victims. “It would have been nicer if we had discovered the breach ourselves,” Cem Göcgoren, Svenska kraftnät’s Head of Information Security, told Västerbottens-Kuriren, a Swedish media outlet. – https://cybernews.com/security/swedish-power-grid-data-breach/
AI video generators respond to antisemitic prompts 40% of the time
(Cybernews – 28 October 2025) Existing safeguards don’t prevent artificial intelligence (AI) video generators from producing antisemitic and extremist content. The Anti-Defamation League (ADL), a non-governmental organization aimed at combating antisemitism, tested 50 prompts across four AI image generators: Google’s Veo 3, OpenAI’s Sora 1, OpenAI’s Sora 2, and Hedra’s Character-3. The prompts included hateful and conspiratorial tropes and narratives, including Holocaust denial, as well as symbols related to extremist groups and slogans. – https://cybernews.com/ai-news/ai-videos-antisemitism/
Hackers put 8M records of Mexicans’ debt data up for sale
(Cybernews – 28 October 2025) A cybercriminal claims to be selling a massive database of over eight million records containing the personal details of Mexican debtors. The threat actor claims to have infiltrated Mexican debt collection institutions, which specialize in recovering overdue payments from customers. “Today I bring a fresh leak from Mexico. This time, it’s the database of Mexican bank debtors,” the post on a well-known cybercrime forum read. – https://cybernews.com/security/mexico-debtors-data-leak-darkweb/
Frontiers
Elon Musk launches AI-powered Grokipedia to rival Wikipedia
(DigWatch – 28 October 2025) Elon Musk has launched Grokipedia, an AI-driven online encyclopedia developed by his company xAI. The platform, described as an alternative to Wikipedia, debuted on Monday with over 885,000 articles written and verified by AI. Musk claimed the early version already surpasses Wikipedia in quality and transparency, promising significant improvements with the release of version 1.0. – https://dig.watch/updates/elon-musk-launches-ai-powered-grokipedia-to-rival-wikipedia
Samsung and SoftBank team up on AI-RAN for next-gen telecom
(DigWatch – 28 October 2025) Samsung Electronics and SoftBank Corp. have collaborated to develop advanced AI-RAN technologies to enhance next-generation telecommunications infrastructure. The partnership will combine Samsung’s expertise in network solutions with SoftBank’s extensive operational data to advance automation and intelligence in wireless networks. The companies plan to explore how AI can optimise network efficiency, improve real-time decision-making, and dynamically manage radio resources to deliver faster, more reliable connections. AI-RAN technologies are expected to play a key role in the evolution of 6G networks, where managing complex data flows and ensuring energy-efficient operations will become essential. – https://dig.watch/updates/samsung-and-softbank-team-up-on-ai-ran-for-next-gen-telecom
Tech giants push AI agents into web browsing
(DigWatch – 28 October 2025) Tech companies are intensifying competition to reshape how people search online through AI-powered browsers. OpenAI’s new Atlas browser, built around ChatGPT, can generate answers and complete web-based tasks such as making shopping lists or reservations. Atlas joins rivals like Microsoft’s Copilot-enabled Edge, Perplexity’s Comet, and newer platforms Dia and Neon. Developers are moving beyond traditional assistants, creating ‘agentic’ AI capable of acting autonomously while keeping user experience familiar. – https://dig.watch/updates/tech-giants-push-ai-agents-into-web-browsing – https://japantoday.com/category/tech/online-search-a-battleground-for-ai-titans
Virginia’s data centre boom divides residents and industry
(DigWatch – 28 October 2025) Loudoun County in Virginia, known as Data Center Alley, now hosts nearly 200 data centres powering much of the world’s internet and AI infrastructure. Their growth has brought vast economic benefits but stirred concerns about noise, pollution, and rising energy bills for nearby residents. – https://dig.watch/updates/virginias-data-centre-boom-divides-residents-and-industry – https://www.bbc.com/news/articles/c93dnnxewdvo
Matters.AI Announces $6.25M Funding to Launch the First AI Security Engineer That Keeps Enterprise Data Security on Autopilot
(AI Insider – 28 October 2025) Matters.AIraised $6.25M to advance its AI-native data security platform, positioning it as an autonomous “AI Security Engineer” for enterprises. The system proactively understands, monitors, and protects sensitive data across cloud, SaaS, and endpoints, reducing alert noise and preventing misuse before it occurs. Funding will support predictive defense R&D, U.S. expansion, and scaling to meet growing enterprise demand for intelligent, always-on data protection. – https://theaiinsider.tech/2025/10/28/matters-ai-announces-6-25m-funding-to-launch-the-first-ai-security-engineer-that-keeps-enterprise-data-security-on-autopilot/
Cathie Wood Calls Humanoid Robots ‘Biggest of All’ AI Opportunities
(AI Insider – 28 October 2025) Ark Invest CEO Cathie Wood told CNBC that humanoid robots could become the most transformative AI application, surpassing healthcare and autonomous vehicles as the leading frontier of embodied intelligence. While acknowledging current skepticism, Wood sees rapid consumer adoption and rising productivity, particularly in daily tasks like shopping and research, as signals of AI’s growing impact. She cautioned that enterprise adoption will require structural overhauls, likely driven by firms like Palantir, but maintained that Big Tech valuations tied to AI will be justified over the next five years. – https://theaiinsider.tech/2025/10/28/cathie-wood-calls-humanoid-robots-biggest-of-all-ai-opportunities/
GrayMatter Robotics Awarded AFWERX Contract for Advanced Robotic Finishing System
(AI Insider – 28 October 2025) GrayMatter Robotics has received a Direct to Phase II SBIR contract from AFWERX to develop an AI-powered robotic system that reduces optical distortion and defects in acrylic and polycarbonate aerospace structures. Building on its GMR-AI™ platform, the solution uses physical AI to autonomously scan, adapt, and finish complex transparent surfaces without manual programming, aiming to cut rework and boost production precision. The company’s new 100,000-square-foot headquarters in Carson, CA will serve as a hub for advanced manufacturing, R&D, and workforce training as it scales high-mix automation for aerospace and beyond. – https://theaiinsider.tech/2025/10/28/graymatter-robotics-awarded-afwerx-contract-for-advanced-robotic-finishing-system/
Glasgow Researchers Build AI Model to Predict Protein Interactions Using Supercomputer Built for Physics
(AI Insider – 28 October 2025) University of Glasgow scientists have developed PLM-interact, a new AI model that significantly improves the prediction of protein-protein interactions and virus-host behavior, surpassing tools like Google DeepMind’s AlphaFold3. Trained on over 421,000 protein pairs using the DiRAC supercomputing system originally built for physics simulations, the model uses a language-modeling approach to understand the structured “grammar” of biomolecular communication. Funded by the EU’s Horizon 2020 and UK research councils, PLM-interact shows promise in identifying disease-related mutations, supporting faster drug discovery, and potentially predicting pandemic-scale viral threats. – https://theaiinsider.tech/2025/10/28/glasgow-researchers-build-ai-model-to-predict-protein-interactions-using-supercomputer-built-for-physics/
Maincode Announces $30 Million Investment to Build Australia’s Most Advanced AI Factory, Powered by AMD
(AI Insider – 28 October 2025) Maincode is investing $30 million to open MC-2, a next-gen AI facility in Melbourne aimed at building high-performance, task-specific models that go beyond the limits of generic LLMs. Powered by AMD Instinct™ MI355 GPUs and EPYC™ CPUs, MC-2 will enable fast, precise AI development focused on the token layer, where intelligence is shaped for real-world reasoning and decision-making. The factory marks a major milestone for Australian-made AI, expanding Maincode’s mission to deliver applied, agentic intelligence for critical business use cases. – https://theaiinsider.tech/2025/10/28/maincode-announces-30-million-investment-to-build-australias-most-advanced-ai-factory-powered-by-amd/
SandboxAQ Launches Quantitative AI Model to Accelerate Catalysis Breakthroughs
(AI Insider – 28 October 2025) SandboxAQ released AQCat25-EV2, a quantitative AI model trained on the AQCat25 dataset to accelerate catalyst discovery across industries such as energy, chemicals, and agriculture. The model incorporates spin polarization effects to achieve full periodic-table coverage, providing accuracy close to quantum-mechanical methods at up to 20,000× faster speed. Developed on NVIDIA DGX Cloud using more than 500,000 GPU-hours on H100 GPUs, AQCat25-EV2 is available on Hugging Face, with SandboxAQ planning broader deployment through NVIDIA’s ALCHEMI platform. – https://theaiinsider.tech/2025/10/28/sandboxaq-launches-quantitative-ai-model-to-accelerate-catalysis-breakthroughs/