A cyber espionage campaign uncovered by Anthropic used the AI model Claude to automate attacks on an unprecedented scale. The company says this marks a turning point in cybersecurity. AI company ...

Yann LeCun accuses Anthropic of regulatory capture. The dispute centers on an AI-driven cyberattack that Anthropic says happened with almost no human oversight and posed a serious cybersecurity threat ...

OpenAI has released a prompting guide for GPT-5.1, its latest model designed to follow instructions with more precision. The guide walks developers through how to update existing workflows and adjust ...

OpenAI is testing a group chat feature for ChatGPT in Japan, South Korea, Taiwan, and New Zealand. Users on Free, Go, Plus, and Pro plans can chat together with other people and ChatGPT in the same ...

AI-faked images and videos - so-called deepfakes - have developed rapidly in recent years. Here, we trace the history and describe the most important milestones. Ad What do all the people in the cover ...

Perplexity has introduced "Comet," an AI-powered browser designed to replace traditional tabs with an interface that uses LLMs to help users navigate the web. With Comet, users can write emails, plan ...

OpenAI is rolling out GPT-5.1, aiming to make ChatGPT more flexible, responsive, and personal. The update brings faster responses, clearer language, and a conversational style that adapts to how users ...

Perplexity has launched "Labs," a new feature for Pro users that can run complex projects like reports, tables, dashboards, or basic web apps. Unlike the faster "Research" mode, Labs works ...

Inception Labs has introduced Mercury, a new series of large language models that use diffusion technology rather than traditional autoregressive processing. The company reports these models can ...

OpenAI has rolled out an updated version of GPT-4o in ChatGPT, extending its knowledge base through June 2024. The refresh aims to provide more current and contextual responses across topics. The ...

Salesforce's new CRMArena-Pro benchmark reveals major challenges for AI agents in business contexts. Even top models like Gemini 2.5 Pro manage just a 58 percent success rate on single turns. When the ...

Microsoft researchers have developed what they're calling a "Large Action Model" (LAM) - an AI that can operate Windows programs on its own. This marks a shift from AIs that just talk about doing ...