Archive.

OpenAI released an updated public policy agenda outlining its positions on AI safety evaluations, youth protection, labor market impacts, and international standards for frontier models, calling for stronger global coordination on AI governance.

The Biden administration announced the U.S. AI Safety Institute Consortium, bringing together more than 200 organizations including OpenAI, Alphabet’s Google, Microsoft, Meta, Apple, Amazon, and Intel to develop guidelines for red‑teaming, capability evaluations, risk management,

Google DeepMind has launched Deep Research Max, an autonomous research agent designed to perform complex literature reviews, data analysis, and report drafting across scientific and business domains. The tool targets enterprise and research customers who need AI systems that can

Chinese AI startup Moonshot AI has released Kimi K2.6, an open‑source coding model optimized for long‑horizon software projects, complex web front‑ends, and large agent swarms. Kimi K2.6 can coordinate hundreds of parallel sub‑agents for tasks like building WebGL shaders and mult

Alibaba Cloud has introduced two new large language models, Qwen 3.6 Max Preview and Qwen 3.6 Claim, with improved agentic coding, stronger world knowledge, and enhanced reasoning across text and multimodal tasks. Qwen 3.6 Claim is released as a fully open-source model aimed at d

OpenAI and Microsoft have expanded their long‑term partnership with a new multibillion‑dollar Azure compute commitment, deepening integration of OpenAI models into Microsoft products and cloud services. The agreement reinforces Azure as OpenAI’s primary cloud provider while givin

The Indian government has launched a national program to train hundreds of thousands of public-sector employees on generative AI tools, automation platforms, and data literacy. Officials say the initiative is designed to improve service delivery while mitigating displacement risk

Accenture released a new report outlining how large clients in insurance, consumer goods, and public sector are using generative AI for claim processing, marketing content, and citizen services. The documented programs include quantified productivity gains and lessons on governan

The UK’s Financial Conduct Authority and Equality and Human Rights Commission have jointly opened a review into how banks and insurers deploy AI in recruitment, promotion, and performance evaluation. Regulators aim to assess risks of algorithmic bias and potential discriminatory

A new briefing from members of the European Parliament highlights concerns that widespread AI deployment is enabling covert workforce reductions, especially in customer service and back-office roles. Lawmakers are calling for stronger transparency rules and worker consultation wh

Google has unveiled Gemini for Science, a suite of AI tools aimed at helping researchers with literature review, code generation for simulations, and automated data analysis. Early pilots in universities and pharma R&D labs suggest productivity gains in experimental planning and

Anthropic has open‑sourced a suite of interpretability and mechanistic analysis tools designed for large transformer language models similar to Claude. The release includes code, example notebooks, and documentation aimed at helping researchers study internal representations and

Anthropic has released Claude 3.7 Sonnet, an updated version of its mid‑tier model that shows higher scores on reasoning, coding, and multilingual benchmarks compared with Claude 3.5 Sonnet, while incorporating refined safety guardrails.

Malaysia has approved and launched Ryt Bank, billed as the country’s first AI-driven digital bank, using AI for credit scoring, fraud detection, and personalized financial products. Regulators say the move is part of a broader strategy to expand financial inclusion through automa

Klarna has integrated with Google’s new Unified Commerce Platform (UCP) to support payments made by AI agents, allowing autonomous shopping and subscription bots to initiate and complete transactions. The collaboration targets retailers and marketplaces preparing for agent-driven

Thailand has been selected as one of the first countries in Asia where OpenAI’s Sora video-generation app is officially launching, reflecting the country’s fast-growing creator economy and digital policy push. The rollout is expected to spur AI use in local media, advertising, an

Microsoft’s AI leader Mustafa Suleyman predicts that routine computer-based tasks in fields such as accounting, legal, and marketing will soon be fully automated, attributing the trend to rapid AI advances.

OpenAI has officially launched ChatGPT for Microsoft Excel, integrating advanced AI capabilities directly into the spreadsheet program. This new feature aims to significantly enhance data analysis, formula generation, and overall productivity for users.

NVIDIA achieved a new record in large language model inference benchmarks, highlighting ongoing hardware performance leadership for AI deployment.

Anthropic opened its Claude Marketplace, providing enterprise customers access to Claude-powered apps and services developed by third parties, including Replit and GitLab.

A protest occurred at OpenAI headquarters as public concern grows over the company’s involvement in military artificial intelligence projects.

OpenAI has officially unveiled GPT-5.4, their latest foundational large language model, promising significant advancements in reasoning, context understanding, and multimodal capabilities. This release marks a new era in AI performance and application potential.

Nvidia committed $100 billion to supply GPUs for OpenAI’s developing data centers, signaling sustained investment in AI infrastructure.

OpenAI has released GPT-5.3 Instant, an upgrade to its most-utilized ChatGPT model, offering more consistently helpful conversational flow and improved response accuracy. The update is now available to ChatGPT users and via API.

OpenAI closed an agreement with the U.S. Department of Defense to supply AI technology for classified environments, sparking a backlash around AI safety and ethical usage; CEO Sam Altman issued statements promising safeguards.

Chinese developer MiniMax released the M2.5 large language model, achieving performance comparable to Claude Opus 4.6 at one-tenth the price. The launch has drawn significant developer interest and global investment attention.

Anthropic CEO Dario Amodei rejected a Pentagon ultimatum to provide guardrail-free access to Claude models for all lawful government purposes, citing safety and accountability concerns.

On February 26, 2026, tensions between AI companies OpenAI and Anthropic continue over strategic directions, including the dropping of Anthropic’s vow to avoid releasing risky AI models amid Pentagon contract pressures.

Anthropic launched Claude Sonnet 4.6, its most capable Sonnet model yet, featuring a 1 million token context window and enhanced coding, reasoning, and agent planning capabilities.

Led by Yoshua Bengio, the February 2026 report assesses general-purpose AI capabilities, risks, and management approaches including Frontier AI Safety Frameworks adopted by leading AI companies, noting imperfect technical safeguards.

Recent state legislation includes Florida’s AI Bill of Rights legislation and Nebraska’s Conversational AI Safety Act, focusing on AI safety protocols, privacy protections, and banning certain AI uses like legal personhood and deepfakes.

OpenAI began rolling out advertisements on ChatGPT, impacting the service’s free usage model and changing the economics of free AI access for its large user base.

Nvidia and Meta formalized a multi-year partnership to create the world’s most advanced AI infrastructure, signaling a massive strategic move in the AI technology sector.

HUMAIN injected $3 billion in xAI’s Series E round just before xAI’s acquisition by SpaceX, converting its investment into SpaceX shares and becoming a major minority stakeholder.

Meta committed to a $135 billion AI investment mostly focused on infrastructure, securing millions of NVIDIA Blackwell AI chips and pledging tens of gigawatts of computing power through 2028.

Alphabet’s Waymo launched its autonomous robotaxi service in Miami, its sixth U.S. city, starting with a 60-square-mile area with plans to expand coverage and add highway and airport routes in 2026.

Google announced the launch of Gemini 3.1 Pro, an upgraded model with over double the reasoning performance of its predecessor Gemini 3 Pro, available in preview via Google AI Studio and Gemini API.

Alibaba introduced Qwen 3.5 on February 16, 2026, a 397 billion parameter hybrid vision-language model supporting 201 languages, emphasizing multilingual capabilities and efficient sparse compute.

Zipline raised $600 million in a recent funding round, highlighting investor interest in AI-powered platforms with applications in physical-world logistics and delivery.

Tech stocks saw a major correction as investors recalibrated expectations on AI disruption’s impact, particularly on software sectors, with Deutsche Bank calling it a market readjustment from overly optimistic valuations.

Amazon announced additional layoffs reallocating workforce towards AI-focused positions amid broader industry shifts in work processes as emphasized by Meta CEO Zuckerberg.

A global survey showed companies expect to spend 1.7% of revenue on AI in 2026, up from 0.8% in 2025. 72% of CEOs now lead AI decisions, signaling deep enterprise adoption.

Meta finalized its acquisition of Manus, valued at over $2 billion, and plans to integrate Manus’s autonomous agents into its advertising and shopping platforms to offer enhanced agentic tools.

Meta and OpenAI are in acquisition talks for OpenClaw, a top autonomous AI agent GitHub project used for automating email, browsing, and file management tasks.