Reliable energy and trusted digital infrastructure are emerging as linked priorities for AI growth. In Egypt, a $650 million project involving Norway’s Scatec and three multilateral development banks pairs one gigawatt of solar power with 200 megawatt-hours of battery storage. It is designed to supply about one-third of Egypt Aluminium’s electricity needs and may reduce emissions. Egypt is also developing local AI, including the Karnak Chat and Loghat applications launched in September 2026, while seeking stronger Arabic-language tools and protections for national data.
At B20 USA 2026 in Washington, Telefónica and other Spanish companies called for stronger networks, predictable rules and public trust to support AI adoption. Europe needs an estimated €200 billion to meet its connectivity targets; the Draghi report counts 34 mobile operator groups in Europe versus three in the United States. Telefónica supports simpler regulation and more investment in digital infrastructure. Separately, a proposed UN framework for AI governance could set global safeguards, with citizens’ assemblies and a UN Parliamentary Assembly suggested as routes for public input and political oversight.
AI’s effects also depend on how people encounter it. In a Penn State experiment with 135 participants, agreement between an AI second opinion and a doctor increased trust in the doctor’s recommendation; disagreement led participants to rate the doctor as lazier. The findings highlight the risk that poorly presented AI opinions could undermine clinicians’ credibility. Nobel Peace Prize laureate and Rappler CEO Maria Ressa likewise urged transparent, accountable AI use, emphasizing media literacy and critical thinking, particularly for young learners, amid concerns about bias, manipulation and job displacement. An essay on AI agents frames them as holders and tellers of human culture, linking that role to ancient Greek ideas of inspiration.
Businesses are applying AI across varied workflows, while keeping people involved. GlobalData says insurers are expanding beyond underwriting into claims, customer service and product development, but legacy systems, regulation, limited data and algorithm needs remain hurdles. In perfumery, AI helps check ingredient rules and find substitutes or scent combinations; IFF estimates its tools free perfumers’ time by 25 to 30 percent. Makers say human intuition and creativity still guide the final fragrance.
New training and infrastructure efforts target the skills and systems needed to put AI to work. The University of St. Thomas has launched a fully online Master of Arts in Collaborative Intelligence, combining AI with critical thinking, communication and teamwork. It requires no prior AI or coding experience, can be completed in as little as seven months, and begins its first term in February 2027; the inaugural cohort starts in spring. For inference, a WEKA and Nebius test found a storage system handled 2.4 times as many requests as a GPU-memory-only setup and delivered a first token within five seconds 97 percent of the time, compared with 44 percent using only HBM. WEKA’s NeuralMesh uses RDMA and NVIDIA GPUDirect Storage to move KV cache data from NVMe to GPU memory and preserve context across sessions and node failures. The platform is commercially available and runs on H100 systems at OCI and B300 systems at Nebius.
Security leaders warn that AI could speed up vulnerability discovery and exploit development enough to revive self-spreading internet worms. Zane Lackey said future attacks could steal API keys used for AI inference and spread ransomware, urging continuous red-team testing and faster defenses. He cited specialized models from Hinoki Security and startups Neo, Keycard and Socket, and warned that companies may need to install updates much faster than in the past. The supplied articles mention NVIDIA through its GPUDirect Storage technology; they do not report a $650 million NVIDIA investment in Scale AI, a 30% stake, or a 97% ownership figure.
Key Takeaways
- Egypt’s $650 million Scatec-backed project combines one gigawatt of solar power with 200 megawatt-hours of storage, targeting about one-third of Egypt Aluminium’s electricity needs.
- Egypt is developing local AI applications Karnak Chat and Loghat and strengthening Arabic-language tools and national-data protections.
- Telefónica says AI growth needs stronger networks and predictable rules; Europe’s connectivity investment need is estimated at €200 billion.
- A Penn State experiment found that AI disagreement can reduce patients’ trust in a doctor, while agreement increased trust in the recommendation.
- Nobel Peace Prize laureate Maria Ressa called for transparent, accountable AI use and stronger media literacy and critical thinking.
- Insurers are using AI in claims, customer service and product development, though legacy systems, regulation and data constraints remain obstacles.
- IFF estimates AI tools free perfumers’ time by 25 to 30 percent; human creativity remains central to the final fragrance.
- The University of St. Thomas’s online Collaborative Intelligence master’s can be completed in as little as seven months, with its first term beginning in February 2027.
- In a WEKA and Nebius test, storage-based inference handled 2.4 times as many requests and delivered a first token within five seconds 97% of the time, versus 44% with HBM alone; the system uses NVIDIA GPUDirect Storage.
- Zane Lackey warned AI could accelerate self-spreading cyberattacks and urged continuous red-team testing and faster patching. The articles provide no substantiated information about Scale AI or a $650 million NVIDIA investment, a 30% stake, or 97% ownership.
Egypt sees energy and AI as drivers of future growth
Egypt could use reliable energy and trusted digital infrastructure to attract investment, expand exports and create jobs. A $650 million project involving Norway’s Scatec and three multilateral development banks will pair one gigawatt of solar power with 200 megawatt-hours of battery storage. It is designed to supply about one-third of Egypt Aluminium’s electricity needs and may lower emissions. Egypt is also working to develop local AI, including the Karnak Chat and Loghat applications launched in September 2026, and to strengthen Arabic-language tools and protect national data.
Telefónica calls for stronger networks and clear rules to compete in AI
At B20 USA 2026 in Washington from 9 to 11 November, CEOE, Telefónica and other Spanish companies discussed economic growth and innovation. The meetings stressed that AI growth needs strong digital networks, stable rules and public trust. The European Commission estimates that Europe needs about €200 billion to meet its connectivity targets, while the Draghi report counts 34 mobile operator groups in Europe compared with three in the United States. Telefónica supports simpler, more predictable rules, investment in digital infrastructure and broader AI use, with safeguards that build trust.
AI agents reflect and retell human culture
The article describes how AI agents can act as storybooks of human culture. It opens with the ancient Greek idea of an enthusiast as someone inspired by the divine, such as the Pythian oracle or Homer calling on a goddess to sing. The article connects this idea to the way AI agents hold and express human knowledge.
AI disagreement can make patients trust their doctor less
In a Penn State experiment, 135 people discussed a stress-management plan with a doctor. When an AI second opinion agreed with the doctor, participants trusted the recommendation more; when it disagreed, they rated the doctor as lazier. The article says patients often welcome AI for convenient tasks and general information, but prefer a human for emotionally difficult conversations. It warns that health systems should think carefully about how they present AI disagreements, since this can damage a doctor’s credibility.
Maria Ressa urges responsible and transparent use of AI
Rappler CEO and Nobel Peace Prize laureate Maria Ressa warned about AI’s risks, especially for young learners. She said AI can be used for good or bad and stressed that media literacy and critical thinking matter. Ressa also said newsrooms should be transparent about how they use AI and accountable for their decisions. She noted concerns including job displacement, bias and manipulation as AI develops and spreads.
Insurers expand AI beyond underwriting despite key hurdles
Insurers are using AI in claims, customer service and product development, as well as underwriting, according to GlobalData. The technology can help improve efficiency and accuracy, but companies still face challenges with legacy systems, regulation, limited data and the need for more advanced algorithms. Many insurers are investing in AI, and its use is expected to grow as they look to improve operations and customer service.
AI gives perfume makers more time to create without replacing them
Perfume companies are using AI to speed up research, check ingredient rules in different countries and find substitute molecules for materials such as civet. At IFF, the technology is estimated to free up 25 to 30 percent of perfumers’ time. IFF tools can analyze an ingredient and suggest others with similar features or possible complementary scents, helping inspire new combinations. Perfumers and perfume makers say human creativity and intuition remain essential, and the tools do not take control of the final fragrance.
UN AI rules should include democratic input and public oversight
A proposed United Nations framework could help address fragmented and ineffective AI governance by setting global safeguards. The article argues that the public should have a role in shaping these rules. It proposes citizens’ assemblies and a UN Parliamentary Assembly as ways to provide public input and political oversight.
University of St. Thomas launches online AI collaboration master’s degree
The University of St. Thomas has launched a fully online Master of Arts in Collaborative Intelligence program for recent graduates and working professionals. The degree combines AI skills with critical thinking, communication and collaboration, and students can complete it in as little as seven months without prior AI or coding experience. Courses cover AI applications, ethics, workplace communication and collaborative problem-solving, with projects that build a professional portfolio. The first term begins in February 2027, and the inaugural cohort starts in spring 2027. December graduates who apply by Jan. 17, 2027, could finish by the end of summer.
New storage approach aims to make AI inference faster and more efficient
AI inference has different needs from AI training, and the article says infrastructure must change to handle large-scale use more efficiently. In a WEKA and Nebius test, a system using WEKA storage handled 2.4 times as many requests as one using only GPU memory. It delivered a first token within five seconds 97 percent of the time, compared with 44 percent using only HBM. WEKA’s NeuralMesh uses technologies including RDMA and NVIDIA GPUDirect Storage to move KV cache data from NVMe to GPU memory, while preserving context across sessions and node failures. WEKA says its platform runs on H100 systems at OCI and B300 systems at Nebius and has been commercially available since November 2025.
Zane Lackey warns AI could bring back fast-spreading internet worms
Zane Lackey warned that AI could help attackers find weaknesses and create exploits faster than companies can patch them, bringing back self-spreading internet worms. He recalled SQLSlammer and Sasser, which spread through remote flaws and sometimes overwhelmed networks with traffic. Lackey said new worms could steal API keys used for AI inference and spread ransomware, so defenders need overlapping detection and faster responses. He urged companies to run ongoing red-team tests with current tools, including specialized models from Hinoki Security, and pointed to Neo, Keycard and Socket as startups building hardened products. He also said AI attacks could force companies to install updates far faster, recalling the Windows XP era when new machines could be compromised before updates finished.
Sources
- How energy and AI can reshape Egypt's growth strategy
- From B20 to Brussels: networks, scale and certainty for competitive AI
- AI agents are self-enacting storybooks of human culture
- When AI disagrees with the doctor, patients trust the doctor less
- Rappler
- Can AI do more than underwrite insurance?
- How master “noses” are embracing AI in perfumery
- Proposed UN regulation of AI needs democratic input and oversight
- University of St. Thomas Launches Master of Arts in Collaborative Intelligence room
- AI inference must free itself from training-induced burdens
- AI Brings Back the Worm Era, Lackey Warns
Comments
Please log in to post a comment.