AI News

Important AI releases, research, company updates, and industry developments selected from trusted sources and presented with clear context.

news

Google invests millions in Mark Zuckerberg’s efforts to create a ‘virtual cell’

Google DeepMind, Meta, and Isomorphic Labs have jointly invested $300 million into Biohub, a nonprofit founded by Mark Zuckerberg and Priscilla Chan, to advance their 'Virtual Biology' initiative. The goal is to create a high-accuracy predictive model of the cell, or 'virtual cell,' enabling researchers to simulate biological experiments digitally and accelerate disease research. This effort is bolstered by over $1 billion in federal funding from the Department of Energy and NIH.

View Details →

news

Building a safer path to autonomous industrial AI

Industrial AI is entering a transformative phase driven by foundation models, physical AI, and agentic systems, enabling automation of complex tasks in critical infrastructure. However, unlike digital-only AI, industrial applications interact directly with physical systems, making safety and reliability paramount. AVEVA's Arti Garg emphasizes that responsible deployment requires new governance frameworks where AI augments rather than replaces human oversight, ensuring guardrails exist for automated actions. Key challenges include integrating disparate data sources (telemetry, logs, engineering docs) for real-time diagnostics and managing the environmental footprint of AI itself. The industry is shifting from predictive analytics to autonomous robots and drones that can operate in hazardous environments, reducing risk to workers while improving efficiency. Realizing this potential demands rethinking business processes, establishing safeguards, and empowering experienced workers to apply their expertise alongside new technologies.

View Details →

news

Surface RTX Spark Dev Box is available for preorder for $5,999

Microsoft has opened preorders for the Surface RTX Spark Dev Box, a high-end AI workstation powered by Nvidia's Arm-based RTX Spark platform. Priced at $5,999 with shipping scheduled for November, the device features a unique 3D-printed aluminum chassis that acts as a heatsink and offers 128GB of unified memory. Designed primarily for developers, it supports running local AI models exceeding 120 billion parameters, making it suitable for prototyping AI agents and on-device coding tools. The unit runs Windows 11 Pro with preinstalled developer environments including Visual Studio Code, GitHub Copilot, and WSL.

View Details →

news

Firmus pulls biggest ASX float since Telstra amid investor doubt about datacentre company

Firmus Technologies has cancelled its planned ASX listing, which was set to be Australia's largest since Telstra in 1997, after failing to generate sufficient investor demand for its AI datacentre business. The company will now seek funding from private markets instead.

View Details →

news

Roundtables: A Conversation With the Creator of AI-Designed Viruses

Stanford PhD candidate Samuel King used a generative AI model to design genetic blueprints for microscopic viruses in 2025. This development marks a significant step toward AI-designed life forms, sparking discussions on the future of biology and bioengineering capabilities.

View Details →

news

Anthropic bans ‘abusive or cruel behavior’ toward Claude

Anthropic has updated its usage policy for Claude to explicitly ban sustained abusive or cruel behavior toward the AI model, framing it as a 'model welfare' issue. The update also strengthens restrictions on election interference, weapon development, and non-consensual surveillance.

View Details →

news

Redefining enterprise intelligence with autonomous AI

The article outlines a critical shift in enterprise AI strategy from simple tool adoption to an 'agentic shift,' where AI acts as an operating model. It argues that structural fragmentation and data silos hinder growth, urging companies to prioritize process redesign over model selection and adopt composable, sovereign data architectures to enable cross-functional intelligence.

View Details →

news

Apple changes full-disk access permissions to curb abuse from AI agents

Apple is updating macOS privacy settings to restrict third-party AI agents from accessing sensitive data like message histories without explicit, granular consent. This move follows a controversy involving Meta's Muse AI agent, which appeared to read user messages despite claims that access was opt-in. The incident highlighted the risks of broad 'full-disk access' permissions, prompting Apple to tighten controls and prevent misuse by powerful AI tools.

View Details →

news

A Flaw in ChatGPT’s Mac App Could Have Let Hackers Grab Sensitive Data

A critical security vulnerability in the macOS version of OpenAI's ChatGPT app allowed attackers with existing malware to hijack the application and access sensitive user data like chat logs and browser sessions. The flaw exploited a trusted script interpreter by spawning it multiple times to bypass signature checks, illustrating significant risks as AI software gains deeper system access.

View Details →

official

Open-sourcing AstaBrief, the fast report-generation model in Asta

Allen AI has open-sourced AstaBrief, a specialized 8B parameter model designed for rapid generation of cited scientific reports. Trained on real researcher queries and optimized via supervised fine-tuning and direct preference optimization, it delivers high-quality synthesis in approximately 51 seconds—significantly faster than proprietary alternatives while maintaining rigorous citation grounding and evidence support.

View Details →