AI in design & creative: 30 real deployments

Launched Style Sketch AI fashion app for outfit assembly from owned clothing, virtual closet management, and virtual try-ons matched to body shape and skin tone in Indonesia, Singapore, and Thailand

Beta AI assistant for slide design, document restructuring, and chart suggestions in Keynote, Pages, and Numbers, built on Apple models and OpenAI

Used Grok Imagine Video 1.5 to produce an AI-generated Odyssey video series, releasing Episode 1 of approximately 7 minutes on October 23, with a full-length AI feature of The Odyssey pledged by end of 2026

Deployed in-house Craft image generation model for all users in the Dream Lab tab, supporting up to 2K resolution across 30 aspect ratios and targeted edits, with API access and MCP connections to ChatGPT and Gemini

optimized Brand Concierge on NVIDIA Nemotron open models and accelerated AI processing in Photoshop, Premiere, and Substance 3D with NVIDIA RTX Spark; accelerated Frame.io embedding generation and processing latency with NVIDIA video decoding and TensorRT

Launching Apple Intelligence features in China in December via iOS 27.3, including Genmoji, notification summaries, writing tools, and object removal in Photos, using a China-specific model trained with Alibaba

Deploys Copilot for AI-generated image editing, with human contractors rating outputs while seeing users' original prompts and uncensored photos

Reported: Reviewers reported repeated exposure to sexualized requests, imagery of young girls, upskirt photos, animal-sacrifice images and fetish content

Launched Playground, a browser-based platform where users describe a game and the system builds a playable 2D or 3D version using Gemini for logic, Nano Banana for images and Lyria for music

Rolling Hy Image 3.5 Preview image model into the Yuanbao chatbot and internal film-editing and design tools

Operated Fable 5 and Mythos 5 AI products in production; pulled offline under a Commerce Department export-control order after a jailbreak was flagged, then restored after shipping a safety classifier

Reported: Commerce Department export-control order pulled Fable 5 and Mythos 5 offline; Anthropic shipped a >99% blocking safety classifier before service was restored

Subscription service deploying Muse AI for image and video generation, Story Restyle and voice effects across Facebook, Instagram and WhatsApp

Reported: 15M subscriptions and trials already active

Online experiment deploying Vibe Design Agents using verbalized sampling for AI-assisted UI generation across 300,000+ tasks

Reported: negative feedback dropped even as code-export gains stayed statistically uncertain

Using autonomous Claude Fable 5.1 agent swarms to generate explorable 3D reconstructions of real places from open data sources including OpenStreetMap and USGS elevation data

Deployed Google Pics AI image editing tool built on Nano Banana into Workspace, supporting object-level edits, text editing, translation, format-aware cropping, and 2K/4K upscaling via Docs, Slides, and pics.new

deployed Grok Imagine Image 2.0 as the default Quality Mode on grok.com/imagine and iOS/Android apps, adding magic-wand region edits, multi-reference generation with up to five input images, smart resize across nine aspect ratios, and workflow templates

Reported: ranks second on both Arena text-to-image (1,320) and image-editing (1,439) leaderboards

Google paused Google Earth's generative image feature after users produced policy-violating synthetic satellite imagery and investigators raised verification concerns.

Meta discontinued an Instagram feature that generated AI images from mentions of public accounts after objections over default use of people's likenesses.

Reported: The feature was reversed days after launch following objections from CAA, SAG-AFTRA and Public Citizen.

Testing mischief 1 image-to-video AI model in production pipelines

AI covers-and-remixes subscription add-on allowing users to create AI versions of opted-in songs, with royalties flowing to original artists and plays counting toward chart positions, targeting Q1 2027 launch

Subscription AI music creation platform trained only on authorized, licensed music with fingerprinting and filtering, with artists and songwriters opting in

Used ByteDance's Jimeng model to generate more than 1.2 million images for the 100-minute theatrical sci-fi film 'Sanxingdui: Future Past' with about 100 creators on a sub-10M-yuan budget

Reported: Film opened nationwide in China with 10,000+ scheduled screenings; opening-day box office projected under 150,000 yuan; production budget under 10M yuan ($1.5M)

Refocusing company on AI tools for music professionals with backing from major labels

Piloting hyperrealistic AI avatars of musicians for live concert performances, enabling simultaneous touring across multiple countries

24-hour live AI news show streamed on X featuring photoreal AI avatars

Reported: cost approximately $50,000 in tokens; avatars visibly struggled to hold coherent conversations

Deploying Seedance AI video and Seedream AI image generation models across TikTok, CapCut and Dreamina

Netflix uses generative AI primarily for post-production work including visual effects, animation and pre-visualization.

Reported: Netflix disclosed that roughly 300 titles used generative AI during 2026.

Operates an AI music platform in closed beta across 20-plus games using adaptive models to reshape licensed songs to match live gameplay; holds 400,000-plus tracks

Roblox embedded a text-to-game generator in its mobile app so users can create basic playable experiences from natural-language prompts.

Reported: The feature enters a consumer product used by roughly 80 million daily active users.

Physics-grounded AI world models for architecture, engineering, and construction deployed on 10 active projects in Texas and California

Five named early adopters use Cadence's agentic platform to automate PCB and advanced-packaging design and multiphysics simulations.

Reported: Cadence claims up to 15x productivity and twice-as-fast time to market; Forvia Hella reduced a component-placement task from four days to about four minutes.