AI News Recap: November 14, 2025
The Week in AI: Fake Bands Hit #1, Moonshot Beats GPT-5, and Self-Rewriting Malware
The Country Star That Doesn’t Exist and the Malware That Rewrites Itself
If you thought the disruption of the entertainment industry was years away, check the Billboard charts where a non-existent country band just claimed the number one spot. This week blurs the lines between reality and simulation, with celebrities licensing their voices for the future and startups offering digital immortality through “personal” AI. While culture shifts, the technical wars are heating up with self-rewriting malware and a new Chinese model that beats American heavyweights at a fraction of the cost.
Table of Contents
👋 Catch up on the Latest Post
🔦 In the Spotlight
🗞️ AI News
🧩 “NeuralBuddies Word Search” Puzzle
👋 Catch up on the latest post …
🔦 In the Spotlight
Billboard’s Top Country Song Is Currently AI Slop
Category: Generative AI & Creativity
🎶 An AI-generated country song titled “Walk My Walk” by a fictional group Breaking Rust is topping Billboard’s “Country Digital Song Sales” chart in the United States.
🤖 Breaking Rust, attributed to a virtual composer Aubierre Rivaldo Taylor, has amassed millions of Spotify listeners and released widely-streamed tracks, despite lacking any real artist biography.
💵 The music industry is accelerating its investment in AI artists, with major record labels signing multimillion-dollar deals, while platforms like Spotify try to manage a surge in AI-generated “slop” content.
Most Worry About AI’s Impact, But Not About Their Jobs
Category: Society & Culture
🤖 Most workers in major economies recognize AI will profoundly affect society and the economy, but most believe their own jobs will remain safe.
🧠 A significant gap exists between awareness of AI’s economic impact and individual action, with many workers lacking AI training or skills for future work.
📊 Despite rapid adoption of AI by businesses globally, workers are uncertain what skills to retrain for and face difficulty integrating AI tools into existing workflows.
Google Flags An AI-Powered Malware Which Rewrites Itself In Real Time
Category: AI Safety & Security
🦠 Google’s Threat Intelligence Group (GTIG) discovered a new AI-driven malware called PROMPTFLUX, which rewrites its code in real time during attacks to evade security systems.
🔄 Unlike traditional malware, PROMPTFLUX can adapt and “learn” on the fly, creating new attack methods and constantly changing its digital fingerprint to avoid detection.
🛡️ The malware is still in testing with no infections detected; Google has disabled related accounts and is developing counter-AI frameworks to defend against future adaptive threats.
Understanding Prompt Injections: A Frontier Security Challenge
Category: AI Safety & Security
🛡️ Prompt Injection Is A Major Social Engineering Threat: Malicious actors can trick conversational AIs by inserting harmful instructions into ordinary content, causing the AI to perform unintended actions or leak sensitive information.
🔐 Multi-Layered Defense Strategies: OpenAI is tackling prompt injection attacks through safety training, automated monitoring, robust security protections, user controls, extensive red-teaming, and bug bounty programs.
🧠 Ongoing Challenge For The Field: Prompt injection remains a frontier research problem; OpenAI continues to invest in making models more robust while educating users and evolving safeguards as AI capabilities grow and risks evolve.
How Moonshot AI Beat GPT-5 & Claude at a Fraction of the Cost
Category: Foundational Models & Architectures
🏆 Moonshot AI’s Kimi K2 Thinking model outperformed OpenAI’s GPT-5 and Anthropic’s Claude Sonnet 4.5 on multiple benchmarks, including Humanity’s Last Exam and BrowseComp, setting new records for reasoning, coding, and agent capabilities.
💰 Kimi K2 was trained at a significantly lower cost (around $4.6 million), with its API shown to be six to ten times cheaper than leading US models, thanks to model innovation and efficient INT4 quantization.
🌏 The open-source release under a Modified MIT License marks a turning point in AI development, narrowing the gap between closed and open models, and challenging US dominance with cost-effective Chinese AI innovation.
Microsoft’s Next Big AI Bet: Building a Humanist Superintelligence
Category: Philosophy & Future of Intelligence
👥 Microsoft is assembling the MAI Superintelligence Team, led by Mustafa Suleyman, to focus on developing “humanist superintelligence”—an advanced AI explicitly designed to serve humanity’s interests.
🏥 The initiative emphasizes practical benefits, targeting applications in healthcare (such as expert-level diagnostics) and renewable energy, with a goal of specialist systems that achieve superhuman performance while remaining controllable.
🔗 Microsoft is setting clear limits to avoid creating “infinitely capable generalist” AI, diverging from rivals and aiming for real-world impact without posing major risks or relying solely on OpenAI models.
Understanding The Nuances Of Human-Like Intelligence
Category: Philosophy & Future of Intelligence
🧠 Phillip Isola, MIT associate professor, explores fundamental similarities among human, animal, and machine intelligence, aiming to discover shared underlying mechanisms.
🖼️ Isola’s research includes representation learning and self-supervised AI, focusing on how both humans and machines form internal models of the world without direct labels, as well as the Platonic Representation Hypothesis suggesting AI models converge on a common reality.
🤝 Integrating machine-like intelligence safely into human society is a major goal, balancing technological advances with insights from cognitive sciences to maximize benefit for humanity.
Top College Degrees That Lead To High-Paying AI Careers
Category: Workforce & Skills
🎓 Math and computer science skills are the most critical foundation for landing high-paying jobs in AI—more important than your specific major.
🛠️ Successful AI professionals combine strong technical fluency with expertise in a chosen domain (such as robotics, healthcare, design, or philosophy) to stand out.
📚 Real-world experience through internships, hands-on projects, and adaptability are essential for career success in the rapidly evolving AI field.
ElevenLabs Strike Deals With Celebs To Create AI Audio
Category: Generative AI & Creativity
🎤 ElevenLabs signed agreements with celebrities, including Michael Caine and Matthew McConaughey, to AI-generate their voices for commercial use and personal content.
🛒 The company launched an AI voice marketplace, making authorized celebrity voices available to brands, with additional personalities like Liza Minnelli and Dr. Maya Angelou to follow.
💸 ElevenLabs is backed by major investors and is part of the broader trend of artists and industry warming to AI-generated audio, while raising ongoing questions about Hollywood’s guardrails and content authenticity.
Anthropic Announces $50 Billion Data Center Plan
Category: Data & Infrastructure
🏗️ Anthropic has committed $50 billion to build custom data centers across Texas and New York, partnering with UK neocloud provider Fluidstack, to meet the growing compute needs for its Claude family of AI models.
🚀 The initiative will bring new data centers online in 2026, supporting Anthropic’s drive for frontier AI development and aiming to maximize efficiency in large-scale workloads.
📊 This massive investment fits Anthropic’s internal revenue projections and is part of a broader wave of AI infrastructure spending, with rivals Meta and the Stargate partnership planning even larger projects.
Immortality Startup Eternos Nabs $10.3M, Pivots To Personal AI That Sounds Like You
Category: Human–AI Interaction & UX
🗣️ Eternos, now rebranded as Uare.ai, developed the Human Life Model (HLM), which uses only an individual’s personal data—not general LLM sources—to create digital replicas capturing unique values, life stories, and decision-making traits.
💸 The company secured $10.3 million in seed funding to launch a platform allowing users to build personal AIs capable of generating content, handling interactions, and executing professional projects.
👤 Uare.ai enables users to fully own and monetize their digital twin, positioning its technology as a scalable tool for creators and professionals while bypassing generic AI and data aggregation models.
Google Maps Releases New AI Tools That Let You Create Interactive Projects
Category: Tools & Platforms
🗺️ Google Maps introduced new AI-powered features, including a builder agent and MCP server, enabling developers and users to generate code for custom interactive map projects and connect AI assistants to Maps data.
🎨 The builder agent can automatically create prototypes for map-based experiences, while a styling agent allows users to personalize maps with tailored color schemes and themes.
🔗 Google added Model Context Protocol (MCP) for AI assistants to query external data sources and launched Contextual View, a low-code component, for answering spatial questions with visualizations in lists, map views, or 3D display formats.
