🚨 Clawdbot AI agent platform exposes hundreds of informations in security breach 🚨
Misconfigured AI agent dashboards exposed API keys, chat histories, and system access.

A recent investigation has uncovered hundreds of internet-facing control interfaces linked to Clawdbot, an open-source AI agent platform that connects large language models with messaging apps and automation tools. These exposed administrative panels allowed outsiders to view configuration data, retrieve API keys, and browse full conversation histories from private chats. Since Clawdbot agents can actively send messages, run tools, and execute commands across services like Telegram, Slack, and Discord, threat actors could impersonate operators, inject rogue messages, and siphon data through trusted integrations. Some deployments even allowed unauthenticated command execution on host systems, in certain cases running with elevated privileges.

🎮 Google ships Project Genie for real-time 3D worlds 🎮
DeepMind's world model generates interactive environments as you explore them.

Google DeepMind has launched Project Genie, an interactive AI world-building prototype powered by its Genie 3 world model, Nano Banana Pro, and Gemini. The tool lets users create and explore 3D environments from text prompts, generating the world in real-time as they navigate rather than producing static images. Users can move through spaces in first or third-person view, with the model maintaining visual consistency when revisiting areas. This marks a significant step toward AI-generated interactive experiences, signaling where consumer AI entertainment could be heading. Currently available to US Google AI Ultra subscribers, sessions are capped at 60 seconds due to compute costs, with each user receiving a dedicated chip during exploration.

🍎 Apple acquires Q.ai in $2B audio AI deal 🍎
Apple's second-largest acquisition ever targets whispered speech and non-verbal commands.

Apple has acquired Q.ai, an Israeli startup specializing in audio-focused AI, in a deal reportedly worth nearly $2 billion—Apple's second-largest acquisition after Beats in 2014. The technology aims to advance features like whispered speech recognition and audio enhancement for products including AirPods and Vision Pro. Q.ai is led by CEO Aviad Maizels, who previously sold PrimeSense to Apple in 2013, which became the foundation for Face ID. Patents indicate the tech could work in headphones or glasses, potentially enabling non-verbal commands to Siri by interpreting facial skin micro-movements. This positions Apple to lead in ambient AI integration across its hardware ecosystem.

INTERESTING TO KNOW

💰 Amazon negotiates $50B investment in OpenAI 💰

Amazon is reportedly negotiating a potential $50 billion investment in OpenAI, with discussions involving CEOs Sam Altman and Andy Jassy. The deal may include access to Amazon's custom AI chips. This represents a major competitive shift given Amazon's prior substantial investments in Anthropic, and would form part of OpenAI's broader $100 billion fundraising round—reshaping AI industry dynamics. The term sheet could be signed in the coming weeks.

⚖️ Anthropic faces new lawsuit over alleged music piracy ⚖️

Anthropic is facing a new lawsuit from music publishers alleging that the AI company's Claude chatbot was trained on copyrighted song lyrics without proper licensing. The complaint, filed by publishers representing major artists, claims Anthropic reproduced and distributed protected works through its AI models. This adds to growing legal pressure on AI companies over training data practices, following similar cases against OpenAI and other players in the industry. The lawsuit seeks damages and an injunction.