Google Veo 3.1: AI Video Generator Adds Audio + 5 New AI Editing Tools

Number eight. VO 3.1’s core engine. The foundation of Google’s update is VO 3.1, their latest video generation model offering enhanced prompt adherence, improved audio visual quality for image-to-video conversion and what the company describes as true-to-life textures with stronger realism than its predecessor. Number seven. Audio for ingredients to video. Previously silent, this feature now generates rich audio while allowing creators to use multiple reference images controlling characters, objects and visual style simultaneously. Number six. Audio for frames to video. Users can simply provide starting and ending frames, and Flow generates seamless video transitions between them, now with synchronized audio for creating more gripping scene transitions.

Number five. Audio for extend feature. The video extension capability, which creates longer shots by building on the final second of existing clips, now includes audio generation. This means that creators can finally produce continuous videos exceeding one minute in length. Number four. The insert tool. This new editing capability allows creators to add various elements into an existing scene, with Flow automatically handling complex lighting and shadow integration. Number three. The remove feature. This tool enables seamless object or character removal with intelligent background reconstruction, eliminating unwanted elements as if they never existed. Number two.

Cross-platform availability. VO 3.1 launches simultaneously across Gemini API for developers, Vertex AI for enterprise customers, and the Gemini app, with new capabilities accessible through both API and enterprise platforms. Number one. Real creative control. The common thread connecting all updates, granular precision. Rather than accepting first take outputs, creators can now iterate, refine and perfect scenes within Flow itself, marking a shift from generation to genuine AI assisted filmmaking. In fact, these experimental features are Google’s response to five months of user feedback, transforming Flow into a comprehensive video creation environment rather than simple generation tool.

But is this the dawn of intelligent, personalized movies, games and music? Or is this the end of an era being washed over by a wave of AI slop? Meanwhile, in robotics, a different kind of AI milestone is unfolding on the ground. A humanoid robot named Ollie is demoing its 31 degrees of freedom simultaneously, relying on fully autonomous whole body manipulation without motion capture data or human teleoperation, achieving what researchers call true embodied AI through four core specifications. Number four. Autonomous perception and execution. Ollie independently senses tennis balls, retrieves them from floor level and deposits its balls into the basket with zero human intervention, which is a leap beyond most tele-operated systems dominating current robotics projects.

Number three. Extended dual arm reach. Ollie achieves a vertical reach of 200 centimeters or just over six and a half feet, with a 180 centimeter horizontal range, which is just under six feet, putting it at the cutting edge of embodied AI handling and real world object manipulation. Number two. Human scale physical design. Standing 165 centimeters tall or just under five and a half feet with 31 degrees of freedom throughout its body and end effectors, Ollie moves at speeds up to five kilometers per hour while managing a maximum payload of three kilograms or 6.5 pounds.

Number one. Zero training dependencies. The greatest progress lies not in what Ollie does, but how it learned. Using no motion capture data and no teleoperation means the system developed manipulation strategies autonomously, departing from data intensive training methods. Altogether, Ollie’s whole body loco manipulation approach suggests embodied AI is moving beyond controlled demonstrations toward genuine autonomous operation. Simultaneously, Microsoft just announced that every Windows 11 PC will become an AI PC through extensive copilot integration, introducing voice activation, visual awareness and autonomous task execution, all centered on three core capabilities, natural interaction, contextual awareness and action taking ability.

Their initiative addresses a fundamental friction point in AI adoption, which is prompting complexity. With 68% of consumers using AI for decision making, Microsoft believes that voice is the solution to tedious text prompting. That’s why now with the new hey, copilot wake word, the AI is activated hands-free as data suggests voice users engage with copilot twice as frequently as text users and copilot vision is expanding globally, offering full desktop and app sharing with content analysis and step-by-step guidance. The feature also includes highlights functionality where users can ask show me how and receive visual instruction within applications, as well as full context awareness in Word, Excel and PowerPoint files, analyzing entire documents beyond visible screen content.

Additionally, Windows Insiders gain access to experimental agentic features, including a reimagined taskbar positioning copilot as a productivity hub with one click access to voice and vision. Copilot actions previews for local file management, enabling task automation like sorting vacation photos through natural language commands, meaning the system functions as a general purpose agent by executing tasks while users focus elsewhere. Next, copilot connectors integrate OneDrive, Outlook and Google services directly into Windows, enabling queries like find my dentist appointment details. Users can even export conversations directly to Word, Excel or PowerPoint, while Windows Settings integration provides direct links through conversational queries.

File Explorer introduces AI actions through partnerships, Manus builds websites from local documents with one click, while Filmora enables video editing. Copilot plus PC users gain zoom integration via click to do. Gaming receives gaming copilot beta on Xbox ally devices, offering recommendations without leaving gameplay. Security anchors the platform through Microsoft’s secure future initiative. Copilot actions remains disabled by default with user controlled activation, full visibility into agent activities and responsible preview testing rollout. But Microsoft also unveiled Dragon Copilot expansions, introducing the first commercially available ambient nursing experience and extensibility features that integrate partner AI directly into clinical workflows.

New extensibility capabilities also allow third-party AI applications and agents to integrate securely within Dragon Copilot’s interface. Anyways, like and subscribe for more of the latest in AI news and check out this video here. [tr:trw].