991 articles
A security incident involving the Cursor AI coding tool has prompted the introduction of 23 new safety guidelines for AI agents. The rules aim to reduce risks tied to autonomous AI systems operating with broad access and permissions. Source: tech-insider.org
Singapore's DBS Bank and payments company Stripe are collaborating to develop payment rails specifically designed for AI agents operating across Asia. The initiative targets the growing need for autonomous systems to handle transactions directly. Source: startupfortune.com
A developer shared how they created a fully functional, production-grade AI agent using a stack built around PHP, cPanel hosting, and Google's Gemini Flash model. The project shows that advanced AI agents can run on relatively simple, widely accessible infrastructure. Source: hackernoon.com
In a talk at Droidcon USA, Giovanni shared insights from training an AI agent to build TV applications. The presentation covers practical lessons learned while developing a coding assistant for this specialized use case. Source: mshale.com
A new security framework proposes three distinct layers to protect systems running autonomous AI agents. The approach follows a defense-in-depth strategy, aiming to limit risks as agentic AI takes on more independent tasks. Source: venturebeat.com
Startup Morphed has launched an AI agent capable of generating an entire creative marketing campaign based on just one text prompt. Source: natlawreview.com
Quantum computing company QuEra is using an AI agent from Anthropic to automate a critical process in operating its quantum systems. The move points to growing interest in applying AI agents to specialized scientific and engineering workflows. Source: thequantuminsider.com
Attackers linked to Russia reportedly exploited Cursor's AI coding agent to gain unauthorized access into six different companies. The incident raises fresh concerns about the security risks posed by AI coding assistants with broad system permissions. Source: startupfortune.com
KPMG and Oracle discuss in a video interview how agentic AI is being used to deploy autonomous digital colleagues within organizations. The conversation covers practical applications and what it takes to roll out such agents at scale. Source: smartbrief.com
Security experts argue that detecting threats from AI agents only at runtime isn't sufficient protection. A more proactive, layered defense approach is needed as autonomous agents take on more tasks. Source: blog.gitguardian.com
Cisco has given each of its roughly 90,000 employees access to their own dedicated AI agent. The move reflects a broader push by large enterprises to embed agentic AI tools directly into everyday workflows. Source: wsj.com
OpenAI is said to be developing an AI agent designed to maintain continuity and memory across sessions, rather than starting fresh each time. Details on the product's scope and release timeline remain unclear. Source: wired.com
A new legal analysis examines a Ninth Circuit ruling addressing who is legally considered to be 'accessing' a website when an AI agent does so on a user's behalf. The case marks one of the first attempts to apply the Computer Fraud and Abuse Act to autonomous AI agents. Source: jdsupra.com
Slack has introduced a new feature called Slack Code, aimed at bringing agentic AI-assisted coding directly into team collaboration workflows. Source: salesforce.com
A new approach proposes granting AI agents increasing independence step by step, rather than full autonomy at once. This graduated model aims to close the trust gap between businesses and autonomous AI systems. Source: aws.amazon.com
SandboxAQ has open-sourced a tool called Switch, designed to let any AI agent connect into any team chat platform. The project aims to simplify integration between AI agents and existing collaboration workflows. Source: prnewswire.com
AccuKnox has launched AgentZ, a platform designed to help enterprises build, deploy, and govern AI agents at scale. The tool targets organizations looking to manage agentic AI operations with more oversight and control. Source: streetinsider.com
AWS and NVIDIA announced plans to deploy an additional 2 million GPUs along with next-generation infrastructure. The buildout is aimed at supporting agentic AI systems and physical AI applications like robotics. Source: nvidianews.nvidia.com
Traditional rule-based safeguards struggle to anticipate every possible action an autonomous AI agent might take. New security approaches aim to catch and stop unexpected or harmful agent behavior that static rules fail to cover. Source: blog.checkpoint.com
Despite growing enthusiasm around AI agents, new research indicates they can inadvertently expose sensitive personal information. The findings raise fresh concerns about privacy risks as agentic AI adoption accelerates. Source: techxplore.com