OpenAI introduces new beta features for ChatGPT Plus users, allowing file upload, multimodal capabilities, and intelligent context-based guessing. ChatGPT can now process various file types for data summarization, Q&A, and data visualization.

Humane's upcoming "Ai Pin" device is a screenless wearable equipped with camera, microphone, sensors, and Qualcomm Snapdragon chip. It will offer AI features driven by a large-scale language model and a unique "Trust Light" privacy indicator. The device may cost up to $1,000 and require a monthly data subscription. Humane aims to revolutionize the consumer computing field beyond Ai Pin.

Voluntary responsible AI commitments by tech giants like Google and Amazon are questioned amid issues like union activities and misalignment in AI models. There is a lack of consensus in value alignment. AI researchers are urged to focus on preventing misalignment and catastrophic risks. Tech companies like Google are criticized for prioritizing profits over ethics.

Google's AI chatbot Bard now offers real-time responses and allows users to interrupt its reply generation. Users can also modify the response style and verify information through the Google logo. Other updates include a new image viewing feature and improved email summarization for Gmail users.

OpenAI launches ChatGPT based on GPT-3.5, making waves in generative AI. Meta's LLaMA model leaked on 4chan, driving open-source AI forward. Google allows access to PaLM via API, encouraging collaboration and research. GPT-3 continues to innovate with upcoming GPT-4, showcasing its wide range of applications in text generation and beyond.

Malicious actors are using AI to enhance cyber attacks, with AI tools credentials like ChatGPT being sold on the dark web. Future trends suggest AI malware may autonomously steal data. BlueVoyant warns of AI aiding in document forgery, increasing the need for cybersecurity professionals.

United Nations establishes advisory body on AI to harness benefits and mitigate risks. Calls for global dialogue on AI governance. Preliminary recommendations to be presented by end of year. Aim to maximize benefits and minimize risks for all humanity. Advisory body includes 39 members from diverse sectors. Other global initiatives include G7's AI Behavior Code and upcoming AI Security Summit in London.

Kuaishou launched "AI Xiao Kuai", an AIGC service in short video scene, providing interactive functions like Q&A and image generation. It is supported by the "Kuaiyi" language model with powerful capabilities. Users can test the service on Kuaishou by @AI Xiao Kuai in comments. This is Kuaishou's move to explore AIGC and enhance user experience in short video business.

Tencent's MixNet model enhances "WenShengTu" by generating images with diverse styles and textures using advanced algorithms. It addresses challenges in semantic understanding, content rationality, and image quality through fine-grained models, spatial perception improvement, and multi-model fusion. Users can input descriptions for desired styles. MixNet is integrated into various Tencent platforms, offering diverse features like interview outlines and AI conversations for internal testing.

Shutterstock introduces new AI capabilities to enhance image editing, including Magic Brush, Variants, and Smart Resizing. AI Image Generator creates visual effects based on user descriptions. AI virtual assistant offers guidance. Creators retain copyright fees post AI-editing, ensuring AI is a tool, not a replacement.

Google Maps is adding new AI features for navigation, search, and user experience. Integrating features like improved visual search, electric vehicle charging info, and immersive views will enhance user interaction and maintain its edge over competitors in the AI-powered mapping space.

AMD prepares to launch the Instinct MI300X to compete with NVIDIA's H100 in the AI field. Partnerships with Moreh and KT aim to enhance ROCm software for AI tasks. The company's focus on software development shows commitment to innovation, with plans to accelerate competition through acquisitions like Nod.ai and Mipsology. ROCm's performance is key in upcoming hardware releases, positioning AMD as a formidable player in AI technology.

Qualcomm and NVIDIA optimize chips for AI. Intel integrates NPU into Meteor Lake. Apple to launch M3 chip MacBook. Microsoft focuses on AI laptops. Apple leads in spatial computing with Vision Pro chip. Meta's Quest 3 powered by Qualcomm.

Forbes introduces Adelaide, a generative AI search platform on Google Cloud providing personalized search for readers. It offers article recommendations and summary answers within Forbes' coverage. Adelaide is Forbes' first generative AI tool, built on Google Cloud, along with Bertie for journalists. Forbes plans to expand Adelaide's knowledge base to their full archive since 1917.

A new framework called "Woodpecker" corrects hallucinations in multimodal large language models without retraining. It enhances accuracy and interpretability in AI systems, addressing a key challenge in AI applications.

Databricks acquired Arcion for over $100 million to simplify data ingestion into the Lakehouse platform. The acquisition will enable seamless, scalable, and cost-effective solutions for enterprise data sources, enhancing security, compliance, and governance. The recent $1.3 billion acquisition of MosaicML further solidifies Databricks' position as a leading platform for AI model development and scalable growth.

A new soft sensor developed by UBC and Honda researchers is highly sensitive and flexible, offering tactile sensitivity for robotics and prosthetics. The sensor, similar to human skin, enables precise, safe interactions with fragile objects like eggs or glass cups. Made of silicone rubber, it uses a weak electric field to detect forces, key for human-robot interaction. The sensor's practicality and scalability make it suitable for widespread applications, advancing the intelligence of machines.

HumanSignal launches Adala, an open-source framework for autonomous data labeling agents, aiming to accelerate data annotation with AI. Adala learns from a ground truth dataset and uses a large language model for tasks. It goes beyond annotation to serve various data processing tasks as a general-purpose agent.

Researchers at the University of Chicago have developed "Nightshade," a tool that can manipulate AI training images without human detection. This tool disrupts text-to-image models, posing a threat to AI content creators. This comes amid legal battles against tech giants for unauthorized use of online content.

Amazon launches new generative AI feature for vendors to create AI-enhanced product images, boosting ad engagement. Similar to other text-to-image generators, it includes themes like "pumpkin spice" for seasonal appeal. Aimed at streamlining branded image creation for sellers and advertisers.