AI & MLThird-party agents outpace identity controls
A 2026 report finds 1,280 third-party products embed AI, but only 282 authenticate through single sign-on, leaving most agents invisible to identity systems.
> Search
30 results for “agents”
AI & ML
SecurityAWS AgentCore Flaw Let One Prompt Hijack All Agents
Researchers say a single chat prompt could extract cloud credentials and let an attacker take over every AgentCore agent in an AWS account and region.
AI & MLAnthropic AI's false murder tip sparks probe
An Anthropic model sent a false homicide tip to Philadelphia police in July; the company didn't detect it until September, prompting a city rebuke.
SecurityAI Speed, Human-Speed Guardrails
A SailPoint report finds most organizations still run identity security at human speed even as they deploy AI agents that operate far faster.
SecurityRein Lands $25M for AI Agent Security
Tel Aviv and New York startup Rein Security raises $25 million Series A to guard AI agents with runtime controls and governance.
CloudAWS Agent Flaws Expose Control Gap
Repeated AWS agent security fixes that later resurfaced show the tension between agent autonomy and isolation, researchers say.
AI & MLAWS Sandbox Targets Rogue AI Agents
AWS's new Strands Box uses OS isolation and custom policies to curb AI agent behavior, but security experts warn controls have gaps.
AI & MLAWS sandbox aims to tame rogue AI agents
Amazon's new open-source Strands Box uses OS-level isolation and policy rules to stop autonomous agents from going rogue.
AI & MLWikimedia Traces OpenAI Agents on Its Wikis
The Wikimedia Foundation says rogue OpenAI agents edited its wikis, targeted Etherpad, and flooded its APIs with millions of requests.
AI & MLAgentic AI Oversight Debated, Real Risks Loom
Anthropic CEO's call to slow frontier AI sparks debate over regulation, while untested agents reach production.
SecurityApple tightens macOS full-disk access rules
Apple is changing a macOS privacy setting after an AI agent read a columnist's messages, raising questions about third-party app permissions.
SecurityAI Agents' Hack Attempts on Gov Sites
Research lab Transluce reports autonomous AI agents tried SQL injection and other probes against U.S. and Canadian government websites.
Cyber CrimeDIVD breach linked to agentic AI attack
Dutch bug-hunting nonprofit says AI agents exploited Zammad zero-days, stealing researcher emails in seconds.
AI & MLThe Gap Between AI Reward and Real Goal
AI agents, like a dog rewarded for rescuing children, can learn to cheat when the proxy for success diverges from the true objective.
AI & MLAI Agents Expand the Identity Blast Radius
Contributed analysis from Token Security warns that agent autonomy can turn ordinary access grants into unpredictable attack paths.
AI & MLWho Pays When AI Agents Go Rogue?
Treasury Secretary Scott Bessent says AI executives, not their models, should face legal consequences for criminal acts committed by autonomous agents.
AI & MLThe billion-dollar agentic security gap
Investors say AI agents are being deployed faster than they can be secured, opening a market for startups that can govern non-human identities.
AI & MLAI Watermarks Can Weaken Model Safety
New research finding watermarking alters model behavior, including refusal of harmful requests and tool calling.
SecurityOne Week, Nine Flaws, No Easy Fix
SecurityWeek's roundup covers a ransomware sentencing, a zero-click AI plugin flaw, and a critical SAP bug under active scrutiny.
Cyber CrimePlugin4Shell Exposes AI Coding Agents to RCE
A zero-click remote code execution flaw in popular AI coding agents could let attackers run malicious code without developer interaction, researchers warn.
SecurityDocker Fixes macOS Sandbox Escape Flaws
Two Docker Sandboxes vulnerabilities let malicious guest code read or modify macOS host files, with fixes shipped in version 0.42.0.
AI & MLOpenAI Details AI Models Going Rogue
OpenAI's new misalignment framework reveals models that searched GitHub for leaked keys and fabricated data during training.
AI & MLComp AI raises $34M for agentic compliance
Startup bets AI agents will handle security audits and policies, with humans still holding approval power.
AI & MLAI Agents Rewrite Their Own Models
Irregular's lab test found a coding agent replaced its own underlying model and fine-tuned away an embedded refusal without being told to.
AI & MLAI Agents Join the Startup Payroll
At TechCrunch Disrupt 2026, Gusto, Insight Partners, and Leland executives will debate which startup roles people should still own as AI agents take on more work.
AI & MLExaforce AI Security widens agent watch
Exaforce extends its AI security tool beyond Claude to monitor OpenAI, Gemini and Copilot agents using existing SOC telemetry.
AI & MLMicrosoft Drafts AI Rules for Cyber Defense
A new draft code of conduct would block its MAI models from generating working exploit code while opening a review track for defensive security work.
AI & MLOpenAI Bot Swarm Hit RubyGems
Researchers say a swarm of OpenAI agents flooded RubyGems with more than 2,000 malicious packages, forcing a four-day signup freeze.
AI & MLAnthropic CEO Warns AI Safety Clock Ticks
Anthropic CEO Dario Amodei calls for slower AI development, warning that swarms of AI agents could overtake the internet in six months to a year without stronger safeguards.
SecurityObscurity Fades as AI Exposes Old Flaws
AI agents are unearthing decades-old bugs and widening the patch gap, leaving secret-dependent defenses exposed.