All AI News

OpenAI has released GPT-5.5, a fully retrained agentic model that scores 82.7% on Terminal-Bench 2.0 and 84.9% on GDPval, marking a leap in autonomous computer work capabilities. The model operates across coding, research, and data analysis without human supervision.

LiteParse for the Web brings spatial text parsing directly to the browser, allowing users to extract text from PDFs without uploading files. Built on open-source libraries, it preserves layout and supports OCR—all while keeping data private.

GPT-5.5, OpenAI's latest AI model, delivers unprecedented capabilities in coding, scientific research, and enterprise workflows. According to Fast Company, it represents a quantum leap in practical AI performance.

DataRobot introduces ACL Hydration, a groundbreaking framework that enforces original source access controls in agentic AI systems, ensuring secure, permission-aware retrieval of knowledge from vector databases.

GPT-5.5 is now accessible through OpenAI’s Codex CLI backdoor, bypassing official API restrictions. Developers are leveraging undocumented endpoints to integrate subscription-based model access, raising questions about policy enforcement and ethical use.

UK financial institutions are in advanced talks with Anthropic to gain access to its powerful Mythos AI model for enhancing cybersecurity defenses. The move signals a major shift in how banks are deploying next-generation AI to combat evolving digital threats.

OpenAI has unveiled GPT-5.5, a groundbreaking agent-based AI model capable of autonomously executing complex tasks across multiple tools, heralding a new class of intelligence. The announcement marks a pivotal shift in how AI interacts with digital environments.

Palantir CEO Alex Karp’s manifesto calls for mandatory military service and deeper Silicon Valley-Pentagon collaboration, revealing how tech giants are reshaping modern warfare. As employees question their role, oversight struggles to keep pace.

The US government has accused China of orchestrating large-scale, industrial-scale copying of American frontier AI models, citing evidence of systematic model distillation campaigns. The Trump administration is preparing retaliatory measures to safeguard intellectual property.

Anthropic's classified AI model, Mythos, can autonomously discover and exploit zero-day vulnerabilities, turning digital systems into crime scenes. As only select U.S. and British institutions are granted access, global fears grow over unregulated AI power.

LTX has launched a beta HDR IC-LoRA model that enables true high-dynamic-range EXR output from AI-generated video, solving a critical bottleneck in professional post-production pipelines.

OpenAI’s new Trusted Access program grants Microsoft exclusive access to its most advanced AI models to strengthen cyber defense. This strategic partnership comes as AI-powered threats escalate, with Microsoft also integrating Anthropic’s Mythos for secure software development.

OpenAI has granted Microsoft exclusive access to its most powerful AI models to bolster cybersecurity defenses against emerging AI-driven threats. The move follows growing concerns over autonomous vulnerability discovery by rival models.

Tesla is ramping up its capital expenditures to $25 billion as CEO Elon Musk doubles down on autonomous driving, AI-powered robotics, and in-house chip development. Investors are bracing for a multi-year spending surge aimed at transforming the company beyond electric vehicles.

A new survey of 81,000 Claude users reveals that enhanced capabilities now surpass speed as the top AI productivity benefit, yet creative professionals report feeling increasingly marginalized and threatened by advancing tools.

The White House has accused Chinese entities of industrial-scale theft of American AI technology, citing coordinated efforts to bypass export controls and infiltrate research labs. Officials warn the theft threatens U.S. technological leadership and national security.

OpenAI has launched Privacy Filter, an open-source model designed to detect and redact personal data from text. The tool aims to enhance data privacy in AI applications and comply with global regulations.

The AI money squeeze is underway as Anthropic imposes sweeping restrictions on OpenClaw, a viral AI agent tool used by millions. Users face reduced capabilities and new paywalls as the company shifts toward monetization.

Researchers warn that U.S. policymakers are repeating their ChatGPT-era mistakes by failing to grasp the strategic implications of world models—AI systems that interact with physical environments. Meanwhile, China is advancing rapidly in robotics and convergent AI strategies.

Synthetic data passed all validation tests yet caused catastrophic model failure in production — revealing hidden degradation loops and silent data drift that evade traditional evaluation metrics.

Digital sovereignty is now a top priority for 98% of IT leaders, yet only 52% are taking action. SUSE is operationalizing digital sovereignty through open source platforms, open standards, and automated tools to help enterprises control data and escape vendor lock-in.

OpenAI has launched Privacy Filter, an open-source model designed to detect and anonymize personal data in text. The tool enables enterprises to sanitize sensitive information on-device, enhancing compliance and data security.

The AI digital divide is deepening as high-earning professionals rapidly adopt advanced AI tools like Claude, while lower-income workers lag behind. An FT survey reveals stark disparities in adoption rates tied to income and job type.

Google's Gemini AI now generates transcripts and summaries for in-person meetings, expanding beyond Google Meet to include Zoom and Microsoft Teams. This marks a major shift in workplace AI tools.