Twice a week, the most important stories in image and video generation - new models, notable research, and meaningful product releases - distilled into a 2-minute read. No hype, no filler.
Free. Unsubscribe any time. No spam, ever.
OpenAI has added virtual try-on capabilities to ChatGPT, allowing users to upload their own photos and see how clothing and accessories might look on them. The update also introduces a Favorites library where users can save products they find through the chat interface. The move extends ChatGPT's growing role as a consumer shopping tool.
Upload a photo, pick a trending dance template, and Groovely animates it into a vertical clip synced to the beat. No camera, choreography, or editing required.
Tavus has unveiled Griffin, a real-time AI video avatar that the company describes as the first "Human Interaction Model." In an internal study, 48 percent of participants could not tell they were speaking with an AI after a one-minute video call - a dramatic leap from the two percent confusion rate seen in prior systems.
Ideogram has released Ideogram 4.5, a model designed to make targeted edits to specific regions of an image while leaving the surrounding content untouched. It outputs natively at 2K resolution and starts at 0.8 cents per image. Partners including Runway, Pika, and Leonardo AI are already integrating the model.
Instagram is rolling out an AI assistant designed to help creators improve their Reels output by analyzing past performance and suggesting future content directions. The tool represents Meta's latest push to embed generative AI more deeply into its creator-facing tools. It is being gradually introduced to users on the platform.
AMD has agreed to acquire World Labs, the spatial intelligence startup founded by AI pioneer Fei-Fei Li, for $8.2 billion. Li will join AMD as Executive Vice President and Chief Scientist, reporting to CEO Lisa Su. The deal gives AMD an in-house AI research unit built around world models that generate and simulate 3D environments.
Adobe is expanding its plugin inside ChatGPT with more hands-on editing controls for images, designs, and PDFs. The update, co-announced alongside OpenAI's DevDay 2026, introduces fine-tuning tools like hue, saturation, and contrast sliders - letting users move between prompting and direct editing rather than relying solely on text instructions. Adobe says further improvements are planned across other platforms, including Claude and Slack.
Runway has launched Runway Ads, a system designed to automate the full lifecycle of performance marketing creative. The product handles generation, localization, publishing, and iteration of ad content, while keeping human approval as a step in the process. It marks Runway's first direct move into the advertising workflow space.
A man in Singapore has been charged after allegedly sharing an AI-generated image depicting a saltwater crocodile in a public waterway. The case highlights how existing laws around false alarms and public mischief are being applied to AI-created content. He faces up to 10 years in prison if convicted.
Runway has joined the OpenAI Marketplace as a launch partner, making its video generation models available alongside tools from other AI labs in a single enterprise workspace. The integration includes Runway's Gen-4.5, alongside Seedance 2.5, GPT Image 2.5, and ElevenLabs V4. Enterprise teams can now access and manage these offerings from one place rather than juggling separate subscriptions.
Marissa Mayer, the former Yahoo CEO, is launching Dazzle, an AI personal assistant that draws entirely on a user's camera roll rather than email or calendar data. The premise is that photos contain a richer, more candid record of a person's life than a inbox ever could. The app uses generative AI to interpret that visual history and surface useful, context-aware assistance.
The World Juggling Federation faced a practical problem for its ESPN broadcasts: how to make an arena look full when it isn't. Founder Jason Garfield turned to Runway's generative video tools to fill empty seats with convincing crowds, while also using the platform to animate leaderboards and produce title graphics for the production.
French beverage brand Bonjour has built a workflow around Runway that lets its 15-person creative team produce multiple animated ad styles from a single script. The approach compresses what once took weeks into a matter of days, with winning variants identified quickly through live Meta testing. It is a practical example of how small creative teams are using generative video tools to run faster creative experiments at scale.
ARRI and HONOR have jointly unveiled the Magic9 Pro Max, a flagship smartphone that brings ARRI's professional color science - including LogC3, ARRI Wide Gamut 3, and ARRI Looks - to a mobile device. The phone pairs dual 200MP cameras with HONOR's proprietary Imaging Chip H1 and supports APV 10-bit 4:2:2 video recording. Pricing and full specs have not yet been announced, with an initial launch planned for China.
AMD is acquiring World Labs, the AI research startup co-founded by Dr. Fei-Fei Li, in an all-stock deal valued at approximately $8.2 billion. The deal will bring Li into AMD as EVP and chief scientist, reporting directly to CEO Lisa Su. World Labs is best known for Marble, a world generation model that creates interactive 3D environments from text prompts.
AMD has agreed to acquire World Labs, the spatial intelligence startup founded by AI pioneer Fei-Fei Li, in an $8.2 billion deal. As part of the agreement, Li will join AMD as executive vice president and chief scientist. The move signals AMD's intent to deepen its presence in AI research and development beyond its established hardware business.
Topview is an AI video platform that has expanded beyond basic clip generation with the launch of its Film Studio - a suite of cinematic tools designed for creators with actual filmmaking backgrounds. The platform offers granular control over performance, camera work, blocking, and visual style, positioning itself as something closer to a production workflow than a prompt-and-reroll generator.
An Italian politician has drawn widespread condemnation after sharing an AI-manipulated image that appeared to undress a female colleague. The incident arrives months after Prime Minister Giorgia Meloni herself shared a deepfake nude of her own likeness to publicly highlight the dangers of such technology. The case is reigniting debate across Europe about legislative responses to non-consensual AI-generated imagery.
Internal documents obtained by 404 Media reveal that human contractors are reviewing prompts and uploaded images submitted by Microsoft Copilot users. The reviewers are reportedly exposed to large volumes of explicit image generation requests, raising questions about privacy, consent, and working conditions in AI data pipelines.
A new build of Google Flow has quietly updated its internal references, swapping out Nano Banana 2.5 Flash for Nano Banana 2.1. The change was spotted by TestingCatalog and suggests Google is iterating on the underlying image model, though no official announcement has accompanied it.
OpenAI's GPT-6 Astra has reached 80 percent accuracy in identifying assembly errors in IKEA furniture from a single photo - up from just 28 percent in November 2025. The jump signals meaningful progress in spatial and visual reasoning, though real-time assembly guidance remains just out of reach for now.
Google has launched Gemini 3.8 Live with Live Avatar, a feature set that brings animated video personas to enterprise AI agents. The update also introduces background tool calls and multilingual speech support spanning 97 languages. Together, these additions push conversational AI agents closer to a more naturalistic, face-to-face interaction model for business use cases.
President Trump shared a doctored photo on Truth Social depicting himself alongside Chinese President Xi Jinping, with experts concluding that an official in the original image appears to have been replaced with Trump aide Natalie Harp. The incident adds to a growing pattern of manipulated images circulating at the highest levels of politics. It raises fresh questions about the role of AI-assisted photo editing in shaping public perception of diplomatic events.
Adobe is bringing Lightroom and Photoshop editing capabilities directly into Google Gemini, allowing users to work with its tools without leaving the AI platform. The move extends Adobe's broader strategy of embedding its software into third-party AI environments, following earlier integrations with ChatGPT, Claude, Slack, and Microsoft Copilot.
Runway's WorldPrompt system powers its Gen World Models 2 (GWM 2), enabling real-time generation of video and audio through persistent context and timed actions. Rather than producing discrete clips, the model maintains a continuous understanding of an environment as it unfolds. The approach marks a notable shift in how world models can be steered interactively.
Google has updated Gemini Live with an animated avatar that lip-syncs and displays facial expressions in real time during conversations. Called Live Avatar, the feature is currently limited to Gemini Enterprise customers and supports 97 languages without degrading video quality. It marks Google's latest step toward giving its AI assistant a visible, expressive presence.
Midjourney has rolled out a set of interface updates on its alpha site, including live style previews that show how a prompt looks across different styles before committing to one. The update also brings thumbnail previews and refinements to the Edit workflow. These changes reflect the team's ongoing work to make the image generation and editing experience more interactive and informed.
Adobe has announced a new plugin bringing its creative tools into Google's Gemini assistant, while also expanding its existing Claude integration with Acrobat support and new interactive editing features. Both updates allow users to access Adobe's suite - including Photoshop, Lightroom, Firefly, and now Acrobat - directly within AI chat interfaces. The changes are live now across all Gemini plans and on Claude's desktop, mobile, and web apps.
Runway has published a breakdown of its Model Router feature, examining how it performs across 250 image-to-video prompts with a focus on balancing output quality against generation cost. The analysis is aimed at developers weighing which router configuration suits their production needs. Results vary depending on the use case, making the comparison a practical reference for teams building at scale.
Google DeepMind has introduced Gemini 3.8 Live, an updated multimodal model paired with a new Live Avatar feature that generates an animated, talking on-screen presence during real-time conversations. The combination allows users to interact with a responsive visual agent rather than a purely voice-based interface. The release marks another step in Google's effort to make AI interactions feel more immediate and embodied.
Sony Pictures Imageworks texture and motion graphics artists detail how Adobe Substance 3D and After Effects formed the backbone of the production pipeline for KPop Demon Hunters, the 2026 Netflix animated film nominated for both Academy Award and Golden Globe honors. From generating over a thousand crowd character variations to crafting intricate Korean embroidery textures, the tools shaped both the scale and the fine detail of the film's distinctive look. Two key artists walk through the speci
Google has introduced Gemini 3.8 Live with Live Avatar, a new capability that adds near real-time visual presence to its conversational AI. The feature pairs Gemini's existing live voice interaction with an animated on-screen avatar that responds as the conversation unfolds. This marks a step toward more visually grounded AI interaction, moving beyond audio-only exchanges.
Meta used its Connect 2026 event to significantly expand the capabilities of Muse, its AI agent, giving it a video avatar presence, a dedicated email address, and the ability to control Mac computers. The updates push Muse closer to functioning as a persistent, autonomous digital assistant rather than a reactive chatbot. Several new hardware devices were also announced alongside the software changes.
Midjourney has pushed a fresh round of updates to its alpha interface at alpha.midjourney.com, with changes shaped largely by community feedback. Among the highlights is a new style preview feature in the sidebar, now powered by faster models. The update continues Midjourney's ongoing effort to refine the web-based alpha experience ahead of a broader release.
Adobe has finalized its acquisition of Topaz Labs, the AI-focused image and video enhancement company, with plans to integrate its technology across Photoshop, Premiere, Adobe Firefly, and other creative tools. Topaz Labs will continue to operate as a standalone brand while its capabilities are woven into Adobe's broader ecosystem. Early integrations are already live in Firefly and Photoshop, with more expected across the product lineup.
Runway has launched a native integration with DaVinci Resolve Studio, allowing editors to access the company's AI video and image generation tools directly within the editing application. Users can generate, restyle, and import clips to their timeline without switching between separate applications. The integration is aimed at reducing friction for professional editors who already work within Resolve's ecosystem.
Adobe's Premiere mobile app is now available on Android, about a year after its initial iPhone launch. The app brings multi-track timeline editing, 4K export, and Firefly-powered AI features to Android devices running version 13 or later with at least 5GB of RAM. Like the iOS version, it is free to download with optional paid upgrades for additional credits and storage.
YouTube is rolling out a set of AI-powered tools inside Creator Studio, targeting the production workflow from scripting through to publishing. The additions include a storytelling assistant, a Gemini-based chat editor for Shorts, and live translation for streams. The features reflect a broader push by the platform to reduce friction at each stage of content creation.
Stanford University is facing scrutiny after using AI tools to alter an official school photograph, replacing the likeness of a Hispanic male student with that of a Black female student. The edit violated the university's own stated policies on image manipulation. The incident raises broader questions about institutional oversight when generative AI tools are used in communications and marketing contexts.
Adobe has finalized its acquisition of Topaz Labs, the company widely recognized for its AI-powered upscaling tools for photos and videos. The deal, first announced in June, is now complete - and Adobe says Topaz Labs will continue operating as its own distinct brand. What this means for existing Topaz users and products remains a close point of interest for the imaging community.
Video creation platform invideo has integrated OpenAI's GPT-6 Astra to sharpen its editing pipeline, reporting a threefold improvement in color correction and grading accuracy. The model also enables the team to generate 50 custom visual effects in a single day. These gains point to how large multimodal models are beginning to take on substantive roles in professional video post-production workflows.
Evoto has released version 8.0, a broad update spanning its desktop app, mobile app, and video product. Key additions include portrait retouching integrated directly into DaVinci Resolve, story-based intelligent culling, and a new preset and asset marketplace. Mobile tethering support also arrives in this release.
A startup called The Biological Computing Co. claims a thin software layer derived from lab-grown neurons can make text-to-video AI run five times faster at 80 percent lower cost. The layer reportedly adds less than 0.1 percent overhead to an undisclosed base model, and the company is seeking an AWS partnership to bring it to market.
ByteDance has launched Dramagic, an AI platform designed to handle every stage of short drama production, from script writing through to video preview. The tool arrives as demand for this format surges in China, where 128,000 short dramas were released in just the first quarter of 2026 - with 95 percent of them reportedly AI-generated.
An American CEO's vacation in Canada came to an abrupt end after she posted an AI-edited photo of herself and three family members wearing "Lake America" branded sweatshirts. The image went viral and triggered a significant public backlash, ultimately forcing the family to leave the country early. The incident highlights how AI-generated and AI-edited imagery can carry real-world consequences when it intersects with charged political or national sentiment.
Higgsfield AI is using GPT-4o and OpenAI's Astra framework to accelerate the development of its video ad creation platform, bringing new features to market in a single day rather than weeks. The company is targeting small businesses that need professional-looking video content without large production budgets. It is an example of how AI coding and reasoning tools are compressing product development cycles in the generative video space.
Alibaba's Qwen team has released Qwen-Image-2.1, an open-weight image generation and editing model that runs on consumer-grade GPUs. At 7 billion parameters, it claims benchmark performance competitive with larger closed models, while supporting transparency layers and up to ten reference images in a single pass. A research license covers public use, though commercial applications require a separate Qwen license.
Runway is working on a system that streams AI-generated video in real time as users type prompts, rather than rendering a finished clip and delivering it after a delay. The approach draws on GWM-1, the company's world model that produces video one frame at a time. Beyond creative applications, Runway sees potential in fields like robotics and autonomous driving.
