Microsoft's Copilot PCs Surge, Handling 2 Trillion Local Inferences Monthly

Key Takeaways

• <ul><li>Microsoft reports over 40 percent of business laptops being built are Copilot+ PCs, with more than 2 trillion local inferences per month.</li><li>Microsoft introduced Microsoft Execution Containers on Windows 11 to secure AI agents with separate permissions and containment.</li><li>Adobe's Costin Ochescu and Mindway AI's Paula Murphy say AI should support human decision-making, not replace it, at Data Talks 2026 in Bucharest.</li><li>Authors Signé and Yee Amezaga call for a global framework with clear principles to govern AI agents and ensure accountability.</li><li>State attorneys general from both parties are making AI regulation a key campaign issue, drawing on past experience suing big tech companies.</li><li>University of Maryland AI guidelines align with Governor Wes Moore's statewide plan focused on protecting residents and children from harmful AI content.</li><li>Author Margo Cohen argues large language models cannot replicate professional human wargaming due to fundamental differences in how AI and humans approach decision-making.</li><li>Boston scientists use AI to record and analyze rat conversations, filtering out human speech to study rodent behavior and help prevent infestations.</li><li>AI supports weather monitoring and disaster preparation efforts for Hurricane Isaias in Central Alabama and Tuscaloosa.</li><li>Rakuten Viki launched The Painted Soul, an AI-assisted animated K-Drama co-produced with South Korean company Imaginus, on its platform with over 127 million registered users.</li></ul>

Romanian leaders face data security challenges from AI agents

Data security is a top concern for Romanian business leaders as AI agents become more autonomous in the workplace. These findings came from Data Talks 2026 in Bucharest on October 8, which brought together 180 professionals from over 60 companies. Professionals see AI agents mainly as tools to improve efficiency and support better business decisions. Preserving human judgment and data protection ranked as leading concerns, while fear of job loss ranked last. Speakers Costin Ochescu of Adobe and Paula Murphy of Mindway AI emphasized that AI should support human decision-making, not replace it.

Microsoft unveils hybrid intelligence and new AI PC security tools

Microsoft opened pre-orders for new Copilot+ PCs shipping October 16 and announced the Surface RTX Spark Dev Box starting at $5,999 MSRP. The company also expects Windows devices built with Nvidia's DGX Station later this year. Microsoft executive Pavan Davuluri said the company is building an operating system for hybrid intelligence using local data, local models and cloud resources. Microsoft introduced Microsoft Execution Containers (MXCs) on Windows 11 to secure AI agents with separate permissions and containment. Over 40 percent of business laptops being built are Copilot+ PCs, and the company reports more than 2 trillion local inferences per month.

AI still cannot replace humans in wargaming, expert argues

Author Margo Cohen argues that large language models will never be able to conduct matrix wargaming the way professional human wargamers do. She points out that LLMs approach challenges as spaces for producing optimal output, while humans approach wargames as testing spaces for survival and complex decision-making. LLMs are prone to breaking rules and acting abnormally, making them unreliable for maintaining secrecy or following wargaming conventions like the Chatham House rule. The article also raises concerns about whether AI can accurately model human forgetfulness and the unpredictability of human decision-making in conflicts. Cohen concludes that the fundamental difference in how computers and humans perceive victory makes AI a poor match for professional wargaming.

AI-assisted animated K-Drama The Painted Soul launches on Rakuten Viki

Rakuten Viki announced the launch of The Painted Soul, an AI-assisted animated K-Drama fantasy romance co-produced by Viki and South Korean company Imaginus. The series follows Yun Hwa, a trainee painter who releases a stranger trapped inside a painting for 50 years. The show was directed by Choi Jinnie, who previously directed Mouse and founded Imaginus in December 2021. Rakuten Viki has over 127 million registered users across more than 190 countries and offers free and subscription viewing options. The announcement was unveiled ahead of the Busan International Film Festival's Asian Contents and Film Market.

Top AI tools for creating training courses in 2026

Colossyan ranks as the best AI tool for creating training courses in 2026, turning documents and briefs into avatar video courses with assessments. Its AI agent Cora drafts every scene, while humans approve each scene before publishing to maintain quality control. Courses export as SCORM packages to existing learning management systems and support localization in over 100 languages. Colossyan meets SOC 2 Type II standards and offers a free Starter tier with Professional plans at $59 per month. The article recommends a five-stage rollout process covering trial, pilot, review, technical check and scaling for teams adopting AI course creation tools.

Boston scientists use AI to study rat conversations

Scientists in Boston are recording the sounds made by the city's rats. They use AI to filter out human speech and focus on rodent calls. By matching these sounds with video of rat behavior, the researchers hope to learn how rats interact and move. This knowledge could help the city prevent infestations before they become a problem.

Authors propose new global rules for AI agents

Landry Signé and Karla Yee Amezaga write about the challenges AI agents create for current governance systems. They argue that existing rules are not enough to handle these new technologies. The authors call for a global framework with clear principles to ensure accountability and transparency. They stress the need for international cooperation and a proactive approach to governing AI agents.

UMD guidelines align with Maryland's new AI framework

University of Maryland leaders say their AI guidelines match many parts of Governor Wes Moore's new statewide plan. The state framework focuses on protecting residents, supporting workers, and shielding children from harmful AI content. The university's Artificial Intelligence Interdisciplinary Institute shares principles like accessibility and social justice with the state. However, the new rules mainly target K-12 students, so college curriculum changes are not expected right away.

State attorneys general campaign on AI regulation

State attorney generals from both parties are making AI regulation a key part of their campaigns. They draw on past experience suing big technology companies to argue they can hold AI companies accountable. Arizona Attorney General Kris Mayes says these companies often do not follow the same rules as everyone else. The elections have turned AI oversight into a major political issue.

AI helps track Hurricane Isaias

Artificial intelligence is being used to help track Hurricane Isaias. The technology supports weather monitoring and disaster preparation efforts. Communities in Central Alabama and Tuscaloosa are also preparing for the storm. Local teams are getting ready to help evacuees and manage the hurricane's impact.

Sources

• Data security remains a major challenge for Romanian business leaders amidst growing autonomy of AI agents
• Microsoft Touts ‘Hybrid Intelligence,’ Agent Security Push And New AI PCs
• Don’t Hold Your Breath on AI-Driven Wargaming
• AI-Assisted Animated K-Drama ‘The Painted Soul’ Sets Rakuten Viki Bow
• 10 top AI tools for creating training courses in 2026
• Boston’s rats are talking. Scientists are listening with AI.
• Shaping global governance of AI agents
• UMD experts compare university guidelines, Maryland’s new state AI framework
• State AGs, Big Tech’s old foes, are running on AI regulation
• Artificial intelligence helping track hurricanes