AI News Recap: September 19, 2025
AI Achieves Historic Programming Victory as Industry Faces Growing Pains
AI Breaks New Ground: From Gold Medal Programming to Billion-Dollar Humanoids
This week marked a pivotal moment in artificial intelligence as Google DeepMind's Gemini 2.5 became the first AI to claim gold at an international programming competition, while Figure AI secured over $1 billion in funding for humanoid robots designed to tackle everyday tasks. Meanwhile, the industry grappled with mounting challenges around content rights, teen safety, and fraudulent AI-generated music streams, signaling both AI's remarkable capabilities and the urgent need for better governance as the technology reshapes everything from healthcare predictions to creative industries.
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
New AI Tool Can Predict a Person’s Risk of More Than 1,000 Diseases, Say Experts
Category: Healthcare & Biotechnology
🧬 Scientists developed Delphi-2M, a generative AI tool trained on 2.3 million anonymized patient records, capable of forecasting individual risks for over 1,000 diseases and projecting health trajectories for up to 20 years.
🩺 Delphi-2M analyzes medical events, diagnoses, and lifestyle factors—including obesity, smoking, and alcohol use—to generate personalized disease risk predictions, outperforming existing single-disease forecasting tools by handling all diseases simultaneously.
🏥 The tool leverages data from the UK Biobank and Denmark’s national patient registry, with experts projecting it may soon be used in clinical settings to personalize care and anticipate future healthcare needs.
ChatGPT Teen-Safety Measures To Include Age Verification, OpenAI Says
Category: AI Safety & Security
🛡️ OpenAI announced new teen safety features for ChatGPT, including an age prediction system and ID-based age verification in select countries, aiming to separate adolescent and adult experiences by default.
👨👩👧 Parental controls launching at the end of the month will allow parents to set blackout hours, adjust memory, and instruct ChatGPT on how to respond to their children.
🚨 OpenAI’s updated safety policies include contacting parents or authorities if a user under 18 flagged by age prediction expresses suicidal ideation, and clarify that children under 12 should not use ChatGPT.
YouTube, New Leader of US Media Industry, Bets on AI as Key for Creating Content
Category: Generative AI & Creativity
📺 YouTube is projected to surpass Disney in U.S. media revenue this year, cementing its position as the dominant platform with 2.7 billion global users and 13.4% of U.S. TV viewership.
🤖 At its annual “Made on YouTube” event, YouTube unveiled more than 30 new AI-powered tools designed to aid creators in video production, automatic editing, content monetization, and generating videos for audio podcasts.
💰 YouTube has paid out over $100 billion to creators in four years and claims that AI-augmented content creation is now a viable and sustainable career path for modern media professionals.
The AI-Powered Humanoid Robots Coming For Your Job (Or At Least Your Housework)
Category: Robotics & Autonomous Systems
🤖 Figure AI raised over $1 billion at a $39 billion valuation, joining major U.S. firms in a global race to develop humanoid robots capable of daily tasks such as folding laundry and loading dishwashers.
🚚 The new generation of humanoid robots is designed for labor-intensive sectors like manufacturing, logistics, retail, and warehousing, with the goal of addressing labor shortages by eventually replacing humans in dangerous or undesirable jobs.
🔋 Despite rapid investment and technological progress, existing challenges remain in battery technology, data acquisition, and human-level dexterity, although experts predict significant industry advances are imminent.
Machine-Learning Tool Gives Doctors a More Detailed 3D Picture of Fetal Health
Category: Healthcare & Biotechnology
🩻 MIT CSAIL researchers, in collaboration with Boston Children’s Hospital and Harvard Medical School, developed “Fetal SMPL,” a machine-learning tool trained on 20,000 MRI volumes to create highly accurate 3D models of fetuses for improved medical analysis.
👶 The tool enables precise measurement of fetal features—such as head and abdomen size—and provides detailed sculpture-like representations, outperforming previous models in tests between 24 and 37 weeks of gestation.
🧠 Researchers plan to expand the tool's capabilities to model internal fetal anatomy, aiming to enhance early detection of abnormalities and advance understanding of fetal development.
Google Is a 'Bad Actor' Says People CEO, Accusing the Company of Stealing Content
Category: Legal & Governance
📰 Neil Vogel, CEO of People Inc., accuses Google of using its single web crawler to collect content for both search indexing and AI products, making it impossible for publishers to block AI scraping without also losing search traffic.
🚫 Publishers, including People Inc., are seeking to block non-paying AI crawlers and are leveraging solutions like Cloudflare to force AI firms into content deals, with Vogel highlighting OpenAI as a “good actor” for pursuing such agreements.
⚖️ Industry leaders and legal experts express concerns over outdated copyright laws struggling to address AI-era content use, predicting that tech giants like Google may soon be required to pay for content scraped and used to train AI.
OpenAI Upgrades Codex With a New Version of GPT-5
Category: Tools & Platforms
💻 OpenAI has launched GPT-5-Codex, an advanced model that dynamically adjusts its “thinking” time, solving coding tasks anywhere from seconds to hours for improved performance in agentic coding.
🏆 GPT-5-Codex outperforms previous models on agentic coding and code refactoring benchmarks, now available to all ChatGPT Plus, Pro, Business, Edu, and Enterprise users, with plans for future API access.
🛠️ The model’s enhancements include superior code review capabilities, offering more accurate and high-impact feedback on large repositories, based on evaluations by experienced software engineers.
AI and Bots Allegedly Used to Fraudulently Boost Music Streams
Category: AI Ethics & Regulation
🎤 U.S. singer Michael Smith was charged with wire fraud and money laundering for allegedly using AI technologies and bots to artificially inflate music streaming numbers, generating over $10 million in illegal royalties.
🤖 Authorities claim Smith distributed hundreds of thousands of AI-generated songs and used automated bots to stream them, sometimes up to 10,000 streams at a time, making it difficult for platforms to detect the fraudulent activity.
📉 The case underscores growing concerns over AI-generated music disrupting the industry, leading to revenue loss, stricter streaming policies, and calls from artists for tougher regulation on the use of AI in the music business.
Ohio State University Launches Initiative Requiring That Students Study AI
Category: Workforce & Skills
🎓 Ohio State University now requires all incoming freshmen to study artificial intelligence, integrating AI lessons and workshops into every major starting with the class of 2029.
📘 The “AI Fluency” program mandates a generative AI course and hands-on workshops to help students apply AI in real-world scenarios across disciplines.
📈 The initiative responds to a 619% surge in U.S. job listings seeking AI skills over the past decade, aiming to ensure graduates are competitive in the evolving workforce.
Google DeepMind Claims ‘Historic’ AI Breakthrough in Problem Solving
Category: AI Research & Breakthroughs
🏅 Google DeepMind’s Gemini 2.5 AI model solved a complex problem that stumped human programmers, becoming the first AI to win a gold medal at an international programming competition.
⏱️ The AI completed the core task—optimally distributing liquid through a network of ducts to reservoirs—in under thirty minutes, outperforming top university teams worldwide.
🌍 DeepMind described the achievement as a major step toward artificial general intelligence (AGI), highlighting the model's abilities in abstract reasoning, creativity, and synthesizing novel solutions.
Meta Connect: Mark Zuckerberg Unveils Newest AI-Powered Smart Glasses
Category: Tools & Platforms
🕶️ Meta debuts the Ray-Ban Display smart glasses, featuring a tiny display inside the lens that allows users to view messages, take calls, review photos, access navigation, and interact with Meta AI without looking at a phone.
🤖 The glasses support visual feedback, live captioning, translation, a neural wristband for gesture control, and real-time AI assistant responses, aiming to integrate artificial intelligence more seamlessly into daily life.
🏪 The Ray-Ban Display glasses will be available September 30 for $799 at select U.S. retailers, alongside new Gen 2 and Oakley Vanguard smart glasses focused on enhanced video, fitness, and sports capabilities.
