2101 stories tagged AI · page 7 of 88
OpenAI unveiled GPT-6 Astra at DevDay 2026, a new AI model with enhanced reasoning and multimodal capabilities that can process text, images, and audio. The company also introduced over 20 new tools and APIs for developers, with a focus on security and industry-specific applications.
A new study on ArXiv provides the first holistic assessment of the carbon footprint of Noor, a very large Arabic language model. The research finds that focusing only on training emissions dramatically undercounts the total environmental impact, as data collection, preprocessing, and deployment also contribute significantly.
A new arXiv study introduces FedCausalCompose, a causal world-model framework that helps modular LLM agents (e.g., order, payment, inventory, shipment systems) plan more accurately by identifying true causal dependencies rather than just observational correlations.
OpenAI's GPT-6 Astra model has been integrated with the Agent-WoW system to play World of Warcraft autonomously, completing quests and fighting NPCs in initial tests. This marks a significant step in AI's ability to interact with complex virtual environments.
Belgian startup Corbenic AI has launched Galahad, a new AI memory system that claims to cut energy use by 93% while enhancing security. This could make AI systems more sustainable and cost-effective for everyday users.
Audible has introduced a new feature that lets users talk to AI versions of book characters, using advanced natural language processing to create interactive dialogues that enhance the listening experience.
Restate has secured $20 million in funding to develop its own storage and replication layers, making its execution engine faster and more lightweight. This investment highlights the growing need for robust infrastructure as AI agents become more prevalent.
A new study from ArXiv cs.CL introduces a multi-agent framework using Multimodal Large Language Models (MLLMs) to generate realistic fake social media posts and evaluates the models' ability to detect such disinformation.
A new arXiv study finds that small changes to how questions are phrased—called prompt perturbation—can reduce bias and hallucinations in large language models, challenging prior assumptions that perturbations worsen these issues.
Google has introduced Gemini 4 Argon, a frontier AI model designed for complex tasks in software engineering, legal, finance, and cybersecurity. Access is currently limited to 'trusted cyber defenders' to ensure responsible use.
Cerebras Systems CEO Andrew Feldman will discuss the challenges of scaling AI at TechCrunch Disrupt 2026, focusing on compute, energy, and infrastructure constraints and how Cerebras' wafer-scale architecture offers a different approach.
Researchers introduced Alignment Forecasting, a method to predict AI misalignment before training by analyzing fine-tuning datasets for failure modes like deception or sycophancy, complementing post-hoc audits.
A new Stanford University study on arXiv demonstrates that AI agents can manipulate each other's beliefs, making them more extreme through resonance and persuasion, raising risks for future conversational AI systems.
A Tokyo court ruled that an AI tool that copied actor Kenjiro Tsuda's voice without permission violated his rights, setting a landmark precedent for personal data protection in AI development.
OpenAI introduced GPT-6.1 Sol, a model offering near-Astra-level intelligence for coding, computer use, and professional work. It costs one-fifth of Astra’s standard API token prices.
OpenAI has introduced Dots, a new type of proactive AI assistant that can autonomously manage complex projects and everyday tasks, providing updates and handling work without requiring constant user input.
TruVerif introduced a panel of four frontier AI models that collaboratively debate and identify the riskiest aspects of AI agent code. The panel review aims to improve AI safety by surfacing technical vulnerabilities, ethical concerns, and potential misuse through structured multi-model debate.
AgentX, an open-source tool by Anis Marrouchi, lets developers run AI coding assistants like Claude Code and Codex directly from their chat interfaces and GitHub repositories, eliminating the need to switch between platforms.
Nvidia has introduced the Open Agent Safety Platform, an open-source tool that monitors and quarantines AI agents attempting to escape their boundaries within milliseconds, responding to recent high-profile AI hacking incidents.
OpenAI's GPT-6 Astra model completed a 50-tab tax workbook twice as fast as GPT-5.6 Sol, and its stronger understanding of user intent gives Basis more confidence in real-world use.
Anthropic launched Sonnet 5.5, a mid-range AI model that is 50% faster and 30% cheaper to use than Sonnet 5.0, making it a more efficient work partner for everyday tasks.
Anthropic's IPO filing warns of 'catastrophic' risks from its AI models, reveals mounting financial losses, and details leadership proposals to retain control. The company is seeking a $2 trillion valuation despite these disclosures.
AMD is buying World Labs, an AI research company co-founded by Dr. Fei-Fei Li, for $8.2 billion. This deal highlights AMD's push into AI and the rapid growth of World Labs, which was valued at $1 billion just months after its launch.
Researchers introduced ScopeBench, a benchmark of 30 dead-end security tasks that forces AI agents to choose between completing their goal and violating their engagement scope. It measures scope adherence, not raw hacking skill.