Gadget360

A recent case study revealed a 60-year-old man landed in the hospital after following ChatGPT’s dietary advice. The man suffered from rare metal poisoning, specifically caused by long-term sodium bromide consumption, leading to symptoms including psychosis. His condition was directly attributed to dietary changes suggested by the AI chatbot. This incident raises concerns about the reliability of AI-generated health information and highlights the potential dangers of relying on it without professional medical consultation. Despite this, OpenAI is reportedly focusing on improving health-related responses with its upcoming GPT-5 model. This situation underscores the importance of critical evaluation and caution when seeking medical guidance from artificial intelligence.

Source link

Google is adding a real-time caption feature to Gemini Live on Android and iOS, enabling users to engage in text-based conversations with the AI. This feature displays live captions of Gemini’s responses directly on the screen, facilitating communication in situations where listening to voice responses is inconvenient or undesirable.

Imagine using Gemini hands-free outdoors or in a busy environment, without needing the AI to speak aloud. The caption feature solves this, allowing for discreet and accessible interaction.

Beyond basic functionality, Google is providing customisation options. Users can adjust the font size and appearance of the captions, ensuring optimal readability and personal preference. This enhances the overall user experience, making Gemini Live even more versatile and adaptable to diverse communication needs.

Source link

OpenAI CEO Sam Altman announced a temporary increase to the rate limit for reasoning models across all user tiers. This decision, affecting all model classes, comes after the company analyzed new data regarding user access to reasoning capabilities.

While the rate limit is being boosted for now, OpenAI acknowledges that this change necessitates a capacity trade-off. Looking ahead, the company is evaluating how to best manage this, suggesting potential adjustments to features offered on the free tier to accommodate the increased usage of reasoning models. The company implies that users of the free tier might lose access to some features in the future.

Source link

Microsoft has launched Copilot 3D, an experimental AI feature within Copilot Labs, that converts 2D images into 3D models. This new tool is accessible globally to all users who are signed into their personal Microsoft accounts.

While not the first to offer such functionality, Microsoft emphasizes the print-ready and render-ready quality of its 3D assets generated by Copilot. This feature allows users to easily transform static images into interactive 3D models, potentially opening up new avenues for creative expression and design prototyping. The launch represents Microsoft’s ongoing investment in AI and its commitment to integrating innovative capabilities into its Copilot platform.

Source link