OpenAI's Continuous-Time Consistency Model (sCM) accelerates multimedia generation by 50× compared to diffusion models, creating high-quality images in 0.1 seconds with two sampling steps. This breakthrough enables real-time AI applications without compromising quality.

Asana launches AI Studio, allowing enterprise teams to build no-code AI assistants that automate repetitive tasks within workflows. Early adopters report significant workload reductions. Available for Enterprise editions, with plans to expand access soon.

Zoom launches AI Companion 2.0, enhancing integration with Zoom Workplace and third-party apps for improved data summarization and tailored responses. It leverages user interactions for personalized services and plans to introduce AI avatars next year, aiming for a smarter workspace experience.

Apple releases iOS 18.2 beta for developers, introducing Apple Intelligence features like Genmoji, Image Playground, AI writing tools, and ChatGPT integration. iPhone 16 users gain Visual Intelligence for camera searches. The macOS Sequoia 15.2 beta is also launched with new AI tools, excluding Genmoji, and geographic support is expanded.

Google has open-sourced SynthID, its text watermarking technology, to improve detection of AI-generated content. Integrated into Gemini, SynthID subtly adjusts token probabilities without affecting text quality, supporting responsible AI development and compliance with emerging regulations.

Microsoft and OpenAI are investing up to $10M to implement AI tools in media outlets, enhancing journalism efficiency. Despite copyright lawsuits, they aim to drive AI integration in the news industry.

Qualcomm and Google partner to develop integrated chipsets and Android-based software for automakers, enabling AI voice assistants and enhanced vehicle intelligence. This collaboration leverages Qualcomm’s automotive chips and Google’s Android ecosystem to drive smarter, automated vehicles.

OpenAI appoints former White House official Aaron Chatterji as Chief Economist to lead research on AI’s economic impacts, including growth, job creation, and infrastructure, leveraging his extensive government and academic expertise.

Midjourney CEO David Holz announced a new AI image editing tool launching next week. It offers extensive manipulation features and can transform any image using depth data. To prevent misuse, Midjourney will enforce restrictions, manual reviews, and metadata tracking, balancing innovation with responsible use.

Cohere upgrades search models with Embed 3, enabling multimodal embeddings for images and text in RAG enterprise search. Supporting 100+ languages and a unified database, it enhances hybrid searches and boosts productivity, positioning Cohere competitively against Google and OpenAI.

Runway AI launched Act-One, enabling users to transfer facial expressions from smartphone captures to AI video characters with the Gen-3 Alpha model. This feature streamlines facial animation for creators while ensuring responsible use through built-in safeguards.

Canva introduces Dream Lab, an AI text-to-image tool powered by Leonardo.ai’s Phoenix model, enhancing image and video editing. Updates include improved Magic Write, interactive charts, expanded asset library, and advanced video tools. Subscription prices for enterprise clients will increase to reflect these AI enhancements.

Anthropic's updated AI models enable Claude to interact with computers like humans by viewing screens, moving the cursor, and typing. Now in public testing, these models show enhanced coding abilities and efficiency, supported by robust safety measures.

Stability AI launches Stable Diffusion 3.5 series, enhancing customization and performance with three models: Large (8B params, 1M pixels), Large Turbo, and Medium for edge devices. Available free for non-commercial use, with licensing for larger businesses.

Genmo released Mochi 1 Preview, an open-source AI video model with improved quality and strict instruction adherence. Available for free trial and download on Hugging Face, it uses a 10B parameter AsymmDiT framework. Genmo raised $28.4M in Series A to develop AGI's creative "right brain."

Huawei unveiled HarmonyOS 5.0, a robust, self-developed IoT operating system integrating the AI-powered Xiaoyi assistant. With enhanced performance, expanded cross-device collaboration, and over 1 billion ecosystem devices, it offers seamless connectivity and advanced features, positioning Huawei as a leader in smart technology innovations.

DeepMind's D-MPC combines diffusion models with Model Predictive Control, outperforming existing planning and RL methods in benchmarks. It adapts in real-time to new dynamics and rewards using multi-step strategies and an SSR planner for robust performance.

Microsoft launches Copilot Studio for SMEs, enabling custom AI agents to automate business processes. Integrated with Dynamics 365, it includes ten new agents. 60% of Fortune 500 companies use Microsoft 365 Copilot, boosting efficiency and reducing costs.

Meta is reintroducing facial recognition on Facebook and Instagram to combat fraudulent ads and assist with account recovery. The technology will detect fake celebrity ads and help users regain access to locked accounts. Privacy safeguards are in place, and the feature will initially exclude Texas, Illinois, the UK, and the EU, aiming for global rollout by 2025.

Midjourney to launch a web-based AI image tool next week, featuring AI-driven editing and object retexturing via text prompts. Initially available to select users through surveys, it includes a dual review system to prevent misuse. The company is also developing video generation models, focusing on balancing speed, cost, and quality.