Welcome back to AI Weekly Updates. Each week we highlight the most important stories from the past week so you can stay ahead without wading through endless headlines. From breakthrough announcements to industry shake-ups, here are the developments that shaped the AI landscape last week.
OpenAI launches o1 reasoning models with PhD-level performance
OpenAI dropped its new o1 series models that can think through complex problems step-by-step. Early tests show PhD-level performance in physics, chemistry, and biology, plus major improvements in math and coding. The catch? They’re slower and pricier than GPT-4o, but sharper on hard problems.
Read more ↗Google rolls out NotebookLM’s Audio Overview feature globally
Google’s NotebookLM can now turn any document into a podcast-style conversation between two AI hosts. Users are going wild turning everything from research papers to personal diaries into surprisingly engaging audio content. It’s like having your own AI radio show.
Read more ↗Adobe faces backlash over new AI training terms
Adobe updated its terms of service, and creatives are not happy. The new language suggests Adobe could use customer content to train AI models, sparking mass cancellations and calls for boycotts. Adobe insists it’s just legal clarification, but trust is hard to rebuild.
Read more ↗Meta’s AI Studio lets anyone create custom chatbots
Meta launched AI Studio, allowing users to build custom AI characters for Instagram and Facebook. Think personalized chatbots for your brand, hobby, or niche interest. Early adopters are already creating everything from fitness coaches to medieval history experts.
Read more ↗Anthropic’s Claude gets real-time web browsing
Claude can now browse the web in real-time during conversations, joining ChatGPT in the actually-current-information club. That means no more “my knowledge cutoff is…” disclaimers for recent events. The AI assistant arms race just got more interesting.
Read more ↗Runway launches Act-One for AI video with expressive actors
Runway’s new Act-One feature can generate video performances using just a driving video and voice recording. The AI maintains facial expressions and emotional nuance that previous tools lost entirely. Hollywood is paying attention, and so should anyone in video content.
Read more ↗Microsoft’s Copilot gets voice mode and vision capabilities
Microsoft’s Copilot can now see your screen and chat via voice, making it feel more like a true AI assistant. The integration across Windows, Office, and Edge is getting tighter, positioning Microsoft as the productivity AI leader.
Read more ↗Perplexity faces publisher lawsuit over content scraping
A coalition of major news publishers filed suit against Perplexity for alleged copyright infringement and “parasitic” content practices. The case could set important precedents for how AI companies can use published content for training and responses.
Read more ↗Apple Intelligence finally starts rolling out in iOS 18.1 beta
Apple’s long-awaited AI features are trickling out to developers. Writing tools, notification summaries, and a smarter Siri are the first wave. It’s not revolutionary yet, but Apple is playing the long game with on-device processing and privacy.
Read more ↗China restricts AI model exports amid growing tech tensions
New Chinese regulations require government approval before AI companies can export models abroad. The move escalates the global AI competition and could fragment the industry into regional blocs. Expect more geopolitical chess moves ahead.
Read more ↗
That concludes this week’s roundup. AI continues to accelerate across every industry and challenge old assumptions about what’s possible. Each Monday we’ll deliver the updates that matter most. Check back next week for another edition of AI Weekly Updates.