Netflix came under fire for using potentially AI-generated or manipulated images in a documentary about Jennifer Pan, sparking concerns about bias. The use of such images, especially for a suspect awaiting retrial, has raised alarm bells about the ethical implications of AI in content production.

Apple is set to unveil the M4 Mac series at a hardware event, featuring the powerful "Donan" chip. The M4 promises top performance, energy efficiency, and AI capabilities, catering to professional users with the Bravo and Hydra series. Updates to iPad lineup expected.

Under new UK legislation, producing sexually explicit deepfake images is now a criminal offense. Without consent, creators face criminal records, fines, and potential imprisonment. The Online Safety Bill also prohibits sharing AI-generated intimate images without consent. The prevalence of illegal deepfake production in the UK is concerning. Anti-deepfake activist Cally Jane Beech calls for an end to violating privacy and dignity through deepfake technology.

Lenovo Tech World showcased Lenovo's latest AI products, including AIPC with Lenovo Xiaotian. Yang Yuanqing highlighted the evolution of PCs to AI Personalized Computers. Lenovo aims to promote AI popularization in every household and various industries, emphasizing the benefits for both individual and enterprise users. They are focused on building a new IT architecture to meet enterprise AI needs.

DingTalk launched the AI Assistant Market with over 200 models in enterprise services, industry apps, finance, legal affairs, education, and entertainment, offering personalized and accurate services. The market marks DingTalk's strategic move towards intelligentization, providing a platform for creating customized AI assistants with low thresholds, enhancing user experience and efficiency.

Intel launched Hala Point, the world's largest neuromorphic computing system using the Loihi 2 processor for brain-inspired AI research. It aims to enhance efficiency and sustainability in AI, advancing artificial intelligence computing.

Hugging Face introduces groundbreaking Idefics2 model for multimodal tasks, surpassing Idefics1 with 8 billion parameters. Integrated with Transformers, it excels in visual question answering, OCR, and dialogue training, offering exceptional versatility and performance.

The Claude 3 series by Anthropic, including Opus, is now on Amazon's AI platform. Opus offers advanced language capabilities and automation for AI applications. Amazon Bedrock hosts the models, with Opus priced at $0.015 per 1,000 input tokens. The Opus model is supported in English, Japanese, and Spanish. Enterprise clients can benefit from Amazon's managed services for Opus' stable operation. Please note Opus is currently limited to the US West (Oregon) region.

Sony partnered with Siemens to launch a new AR/VR headset at CES 2024, focusing on industrial applications. Siemens will integrate the headset with NX immersive design software for 3D design and collaboration. Automotive industry is utilizing metaverse for virtual prototyping and training. Siemens sees strong demand for industrial metaverse products, expecting growth in coming years.

Knowledge management startup Stravito AB has launched a generative AI assistant, Stravito Assistant, for data-driven business decisions. The AI tool offers a conversational approach, guiding users through in-depth research with footnotes and location-based insights. The system prioritizes the latest data and flags potential conflicts. It aims to streamline workflows, save time, and provide actionable insights for faster decision-making.

Adobe has launched Acrobat AI Assistant to reduce search time and enhance document understanding. Available on desktop, web, and mobile, it offers document parsing, Q&A, summaries, and citations. Adobe also integrates AI into web browsers for PDF support and plans to expand AI capabilities for copywriting and editing tools. Firefly Services and Custom Models for enterprises to tailor generative AI models to meet unique needs.

In AI, torchtune, a PyTorch library, simplifies fine-tuning large language models on various GPUs. It offers a comprehensive workflow from data prep to model evaluation, with optimized memory management for 24GB GPUs. Integrated with tools like Hugging Face Hub and PyTorch FSDP, torchtune supports popular LLM models and plans to enhance features, supporting up to 70 billion parameters. This release signifies PyTorch's commitment to AI innovation.

OpenAI introduces new Batch API for efficient asynchronous task processing, compatible with GPT-3.5 and GPT-4. Offers cost control, file management, but lacks streaming media support. Industry experts see it as a significant advancement in asynchronous processing, aiding AI integration in various fields.

Tongyi Qianwen open-sourced CodeQwen 1.5, an advanced code expert model based on the Qwen language model with a parameter scale of up to 7 billion. It excels in code generation, modification, multiple programming languages, and SQL querying - showing impressive performance in various evaluations. The open-source initiative aims to enhance developers' efficiency and promote innovation in the field of code intelligence.

Google is testing a new feature in Chrome Canary that allows users to chat directly with the AI chatbot Gemini. This move shows Google's commitment to AI innovation and improving user experience. It is still in the testing phase and may not be stable yet, but it brings a more intelligent browsing experience.

Kunlun Wey announces significant progress in TianGong series: TianGong 3.0 model improves with 400 billion parameters, enhancing performance in various fields. TianGong SkyMusic public beta opens, showcasing advanced AI capabilities in music composition and audio quality for immersive user experience. Kunlun Wey aims to innovate and provide smarter AI products for users.

AMD launches Ryzen Pro 8040 series mobile processors and Ryzen Pro 8000 series desktop processors using advanced 4-nanometer technology for high performance and efficiency. Ryzen Pro integrates AI tech, enhancing speed by 72% over Intel in AI applications. Ryzen Pro also offers top-notch security features. AMD collaborates with software vendors for optimized AI experiences. AMD faces competition from NVIDIA and Intel in the AI PC market.

Amazon Music is testing Maestro, an AI tool for personalized playlist creation based on text prompts, similar to Spotify's feature. Users can save and share playlists created by Maestro, available exclusively on Amazon Music app. Subscription users enjoy full playlist access, while free users can preview first. Maestro's accuracy is being improved with user feedback, reflecting Amazon's overall AI integration efforts across services like Rufus chatbot.

AI startup Reka launches Core, a cutting-edge multimodal language model with 21 billion parameters. Trained on NVIDIA H100s, it excels in tasks involving text, images, videos, and audio. With advanced reasoning and multimodal understanding, Core offers a 128K context window, multilingual support, and top-notch code generation. Its flexibility and capabilities cater to diverse industry applications. Compare with GPT-4 and Claude 3 Opus on Reka's webpage.

Adobe previewed new generative AI features powered by Firefly AI in Premiere Pro. It integrates third-party models like Sora and Runway, enabling seamless clip extension, object manipulation, and text-to-video generation. Ashley Still, from Adobe, highlighted how these innovations reshape video creation workflows. Adobe emphasizes transparency in content credentials. AI-driven audio tools in Premiere Pro will be available in May.