Global AI Weekly
Issue number: 92 | Tuesday, April 8, 2025
Highlights
OpenAI's models 'memorized' copyrighted content, new study suggests
A recent study suggests that OpenAI's AI models may have been trained on copyrighted material, raising concerns about how these models handle and reproduce protected content. Researchers found evidence that some models can generate text that closely mirrors portions of copyrighted works, pointing to potential ethical and legal implications for AI training practices. This finding reignites debate around transparency and data usage in AI development.
techcrunch.com
OpenAI just made ChatGPT Plus free for millions of college students — and it's a brilliant competitive move against Anthropic
OpenAI is giving college students free access to ChatGPT Plus right before finals, a savvy move in the heated race for dominance in the booming $80 billion education AI market. This initiative not only provides students with advanced study tools but also intensifies competition with Anthropic's Claude and other AI rivals. By targeting students, OpenAI strengthens its foothold in education while showcasing the practical value of its platform.
venturebeat.com
Amazon unveils Nova Act, an AI agent that can control a web browser
Amazon has introduced Nova Act, an AI model integrated into Alexa+ that enables developers to create applications capable of performing tasks autonomously. This advanced agent can navigate and control web browsers, opening up new possibilities for building more dynamic and functional applications. The innovation signals Amazon's continued push into AI-driven solutions for both developers and consumers.
techcrunch.comResearch
WHAMM! Real-time world modelling of interactive environments
Microsoft Research introduces an interactive real-time gameplay experience in Copilot Labs, utilizing the advanced Muse family of world models. Building on the progress made with the World and Human Action Model (WHAM) published earlier this year, the team highlights the latest innovations driving this technology. The blog explores how these advancements enhance the modeling of interactive environments, pushing the boundaries of real-time simulation and gameplay.
microsoft.com
Affordable AI Assistants with Knowledge Graph of Thoughts
This paper introduces a novel approach to building cost-effective AI assistants by utilizing a "Knowledge Graph of Thoughts" (KoT). KoT functions as an organized structure that captures and connects the AI's thoughts, enabling more efficient reasoning and decision-making. The method aims to significantly reduce computational costs while maintaining strong performance, making AI assistance more accessible and practical for everyday use.
arxiv.orgVideo
Humans in AI - Noelle Russell
In this episode of Humans in AI, Noelle Russell, founder of the AI Leadership Institute, discusses the transformative impact of AI on industries like customer service and education. She also shares a glimpse of her life in Weston, Florida, discussing how AI influences small companies and her family of six children. Her insights highlight the practical ways AI is shaping the future beyond traditional tech spaces.
youtube.com
Humans in AI - Tori Tompkins
In this episode of Humans in AI, AI consultant Tori Tompkins from Portsmouth, UK, shares her insights on the future of artificial intelligence, with a focus on small language models and their practical uses at the Edge. The discussion covers key topics like scalability, traceability, and the evaluation of AI systems, offering both enthusiasts and newcomers a deeper understanding of the evolving AI landscape. Tune in for a fun and informative look at the potential of AI in accessible and innovative ways!
youtube.comArticles
15 milestones that shaped Microsoft's vision for AI
Microsoft's journey over the past 50 years has been marked by key technological achievements that paved the way for today's advancements in artificial intelligence. From transformative software developments to groundbreaking tools and platforms, each milestone played a vital role in shaping the company’s AI vision and capabilities. This retrospective highlights 15 pivotal moments that showcase how Microsoft has driven innovation and strived to redefine the role of technology in modern society.
news.microsoft.com
OpenAI disables video gen for certain Sora users as capacity challenges continue
OpenAI has temporarily disabled video generation for some Sora users as it works to address ongoing capacity challenges. The viral image generation feature, introduced in late March, significantly increased demand, overwhelming the platform's resources. OpenAI continues to manage the strain while ensuring a smoother experience for its users.
techcrunch.com
Midjourney v7 launches with voice prompting and faster draft mode — why is it getting mixed reviews?
Midjourney v7 is here, introducing voice prompting and a faster draft mode to enhance the AI experience. While these updates aim to make the platform more intuitive and efficient, the reception has been mixed, with some users praising the innovations while others point out areas still in need of improvement. The discussion reflects growing expectations for smoother integration and user-friendly advancements in AI tools.
venturebeat.comUpcoming Events
Global AI Bootcamp 2025
The Global AI Bootcamp 2025 first month featured over 120 bootcamps all over the world reaching thousands of people, forming new connections, and helping people make sense of AI. There are still over 48 bootcamps happening over the course of the next couple weeks, so if you have not attended a bootcamp yet, consider joining one near you. There is still time! The Global AI bootcamps bring AI enthusiasts and developers together for a series of workshops, sessions, and discussions this March and April. Hosted by local chapters of the Global AI Community, this worldwide event offers hands-on learning opportunities and insights into the latest in artificial intelligence. Join others from around the globe to explore and grow in this exciting field!
globalai.communityCode
ReAct agent from scratch with Gemini 2.5 and LangGraph
Learn how to create a ReAct agent step-by-step using Gemini 2.5 and LangGraph. This guide walks you through the process of building a functional agent, combining the strengths of Gemini’s advanced features with LangGraph's streamlined design. Perfect for those looking to enhance their AI development skills with practical tools and techniques.
philschmid.de
GitHub - tadata-org/fastapi_mcp: A zero-configuration tool for automatically exposing FastAPI endpoints as Model Context Protocol (MCP) tools.
fastapi_mcp is a user-friendly tool designed to seamlessly transform your FastAPI endpoints into Model Context Protocol (MCP) tools without requiring any configuration. This makes it hassle-free to integrate and leverage FastAPI for MCP-related tasks. It's a straightforward solution for developers looking to streamline their workflows while maintaining robust API functionality.
github.com
Semantic Kernel Cookbook
The Semantic Kernel Cookbook is a comprehensive guide that offers practical examples and solutions for working with Semantic Kernel. It provides hands-on recipes to help users easily implement, customize, and optimize their semantic AI applications. Perfect for developers and enthusiasts, this resource makes it simpler to harness the power of the Semantic Kernel framework.
sphenry.github.ioPodcast
Everyday AI Podcast – An AI and ChatGPT Podcast
The Everyday AI Podcast, hosted by Jordan Wilson, is a daily livestream and podcast designed to help everyday professionals enhance their careers using AI. With 20 years of marketing technology expertise, Jordan provides actionable insights on AI tools, trends, and practical tips to simplify work, boost productivity, and stay ahead. From covering advancements by tech giants like Microsoft and Google to exploring tools such as ChatGPT and Midjourney, this podcast offers a wide range of topics to make AI accessible and impactful in your daily routine.
open.spotify.com