AI Update, April 11, 2025: AI News and Views From the Past Week

Catch up on select AI news and developments from the past week. Microsoft celebrates its 50th anniversary and unveils major upgrades for Copilot, positioning it as a cross-platform AI assistant for work and consumer use.

Microsoft expands Copilot Vision to desktop and mobile devices, introducing new interactive formats for customer engagement. Copilot gains research and podcast tools seen in Google's Gemini, making it more useful for campaign planning and multi-format storytelling.

OpenAI's models are found to have memorized copyrighted content, highlighting the need for transparency in AI training datasets. Meta launches Llama 4, its most advanced multimodal AI system to date, while Midjourney's new image model offers faster iteration cycles and improved realism. Verizon reports a 40% increase in sales with the help of Google-powered AI. Amazon adds AI-generated recap feature to Kindle devices, showcasing new personalization techniques for user retention. Google's NotebookLM now supports web-based source discovery, making it easier for marketers to gather and organize research. DeepSeek unveils a new reasoning approach to improve LLM performance. Amazon upgrades its Nova Reel video generation model to produce two-minute videos. Nvidia releases a dense LLM with fewer parameters than DeepSeek R1. xAI hires safety teams to monitor its chatbot, highlighting the importance of robust content moderation. OpenAI's ChatGPT gains a memory feature to track user conversations, with some raising privacy concerns.