#AI agents

OpenAI says it has closed a $122 billion funding round at an $852 billion valuation and unveiled a ChatGPT Super App that combines ChatGPT, Codex, web search, and broader agentic features into one product.

Researchers from four US universities have introduced MetaClaw, a framework designed to let AI agents learn from failed tasks and retrain during idle periods detected through user signals such as calendar activity and ke

As AI agents begin handling travel bookings, purchases, and negotiations autonomously, the quality of their underlying information and contextual awareness is emerging as the critical constraint on whether they can be trusted with real decisions.

A new MIT Technology Review roundup captures a moment when artificial intelligence is no longer just a consumer technology story, but a flashpoint for war, protest, platform backlash, and agent culture.

OpenAI has declared its new North Star: building a fully automated AI research system capable of tackling large scientific problems independently, with an autonomous research intern prototype due by September and a complete multi-agent system targeted for 2028.

OpenClaw, a Chinese AI tool that autonomously controls devices and completes tasks, has sparked a cottage industry of installation services, preconfigured hardware, and a new class of AI entrepreneurs.

OpenAI equips its Responses API with shell access, file management, and sandboxed containers to run secure autonomous AI agents.

OpenAI publishes its design principles for protecting AI agents against prompt injection and social engineering in agentic workflows.

As AI agents take over routine tasks, a new divide is emerging in the workplace: those who leverage agents to multiply their output and those whose roles are being automated away entirely.

A new generation of retrieval-augmented generation replaces static pipelines with autonomous control loops that reason about what to fetch and when.

Perplexity has launched Perplexity Computer, a $200-per-month agentic system that combines AI models from Anthropic, Google, xAI, and OpenAI into a single platform designed to carry out complex workflows independently. The product represents a new approach to AI orchestration.

Amid widespread predictions that AI agents will decimate the traditional software-as-a-service business model, Salesforce CEO Marc Benioff is pushing back, arguing that the SaaS industry has weathered multiple existential threats before. His confidence comes as Salesforce bets heavily on its own AI agent platform.

Anthropic has acquired Vercept, a startup specializing in AI agents that can operate computers autonomously, shortly after Meta recruited one of Vercept's co-founders. The deal signals intensifying competition among AI labs to build agents capable of performing real-world digital tasks.

Atlassian's latest Jira update allows organizations to assign tasks to AI agents alongside human team members. The feature treats autonomous software agents as first-class participants in project workflows, reflecting the rapid normalization of AI in enterprise operations.

A new massively multiplayer online game is designed with no human players at all. Every participant is an AI agent, creating an unprecedented digital ecosystem.

Salesforce transforms Slackbot from a notification tool into a full AI agent powered by Anthropic's Claude, directly challenging Microsoft and Google.

Anthropic launches Cowork, extending Claude Code's agentic capabilities to non-technical users through a folder-based desktop interface on macOS.