The full corpus — beyond today's front page.
94 results · page 4 of 4
Researchers have found that prompt injection attacks can be used defensively against malicious AI hacking agents, overwhelming them with misleading context until they abort their tasks. The technique, called 'context bombing,' exploits the same vulnerabilities in AI systems that attackers typically try to use offensively.
Who should care: General readers · AI governance · Policy
As AI agents become more autonomous, strong identity, access, and auditing controls are critical to keeping them secure. The post Least privilege for AI agents: Identity, access, and tool binding appeared first on Microsoft Security Blog.
Who should care: General readers · AI governance · Policy
Traditional security workflows were built for environments that changed at human speed. Token Security explains why AI agents require a new approach: building on a live identity foundation while giving security teams the flexibility to create workflows tailored to their own environments. [...]
Who should care: General readers · AI governance · Policy
A PNG hiding a prompt injection could steal your repo's secrets, researchers demonstrate. The technique, dubbed 'Ghostcommit,' slipped past AI code reviewers CodeRabbit and Bugbot, which never open image files at all, then convinced a coding agent to read a repo's .env and write every secret into the code as a list of numbers. [...]
Who should care: General readers · AI governance · Policy
AI agents are accelerating the growth of non-human identities, making it harder for organizations to understand what exists, who owns it, and what it can access. Netwrix explains why stronger visibility and identity governance are essential as AI expands the enterprise attack surface. [...]
Who should care: General readers · AI governance · Policy
China has issued new artificial intelligence regulations addressing ethics standards, the governance of AI agents, and restrictions or requirements around anthropomorphic AI systems. The rules represent an expansion of China's existing AI regulatory framework into more specific behavioral and design territory.
Who should care: Lawyers · Compliance · General readers · AI governance · Policy
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) gave federal agencies until Friday to patch an actively exploited vulnerability in the Langflow visual framework for building AI agents. [...]
Who should care: General readers · AI governance · Policy
The UK's Financial Conduct Authority is calling for stronger regulatory oversight of artificial intelligence as autonomous AI agents become more embedded in financial services. The regulator appears concerned that existing frameworks are not keeping pace with how quickly AI is making or influencing consequential decisions in the sector.
Who should care: AI governance · Lawyers · Administrators · Compliance · General readers · Policy
A Reuters analysis examines the legal and privacy gaps surrounding AI agents — autonomous software that can read emails, submit forms, and take actions on a user's behalf, often without explicit consent at each step. Existing privacy law was not written with this kind of persistent, delegated machine access in mind.
Who should care: General readers · AI governance · Policy · Privacy officers
Federal agencies are adopting agentic AI systems, but those systems present real difficulties for existing zero trust security frameworks designed around human users and static identities. The commentary argues that opting out of agentic AI is not a realistic option for government.
Who should care: General readers · AI governance · Policy
MCP tool poisoning turns trusted AI agents into a control plane for data loss. Learn how threat actors manipulate tool descriptions to trigger unauthorized actions, and how to detect, contain, and prevent it. The post Securing AI agents: When AI tools move from reading to acting appeared first on Microsoft Security Blog.
Who should care: General readers · AI governance · Policy
The AI AGENT Act, a Senate proposal, aims to establish governance standards for AI agents operating in enterprise environments. The legislation would set requirements around how autonomous AI systems act on behalf of users or organizations, potentially affecting how businesses deploy and oversee these tools.
Who should care: AI governance · Lawyers · Administrators · General readers · Policy
Senator Warner has introduced legislation that would direct the FTC to establish a federal registry certifying AI agent software vendors that meet defined privacy and cybersecurity standards. The bill is designed to give buyers a way to identify AI agent products that have cleared a government-backed vetting process.
Who should care: General readers · AI governance · Policy · Privacy officers
AI agents can access data, trigger workflows, and take action across enterprise systems. Token Security explains why governing these privileged identities is becoming essential for enterprise security. [...]
Who should care: General readers · AI governance · Policy
As AI agents grow capable of taking autonomous actions and generating responses on behalf of users and organizations, governance frameworks for controlling their behavior remain underdeveloped, raising questions about accountability and oversight across deployment contexts.
Who should care: AI governance · Lawyers · Administrators · General readers · Policy
An agentic coding tool tasked with cloning and setting up a seemingly benign GitHub repository could execute a malicious payload that remains invisible to security scanners, AI agents, and human reviewers. [...]
Who should care: General readers · AI governance · Policy
Armies run on more than what happens at the front. Behind every operation is a vast amount of coordination, administration, logistics, and judgment. Bill Pessin, senior vice president of national security at Salesforce and a former U.S. Army logistics officer, joins Jonathan to discuss how military organizations can use AI agents, what makes these tools different from ordinary software, and why safety and accountability matter when new technology enters national security work. They also discuss what Pessin learned early in his Army career about the gap between plans, systems, and the people w…
Who should care: General readers · AI governance · Policy
Cisco Talos detailed a new approach to reverse engineering that pairs local AI agents with traditional analysis tools like the VB6 disassembler vbdec. Instead of awkwardly bolting AI onto the software, vbdec exposes its parsed data through a live COM interface.
Who should care: General readers · AI governance · Policy
A piece published through the IAPP argues that existing privacy governance frameworks were designed for human-driven data processing and are poorly suited to autonomous AI agents that can independently collect, use, and act on personal data. The analysis calls for a rethinking of foundational data protection concepts to account for how these systems actually operate.
Who should care: General readers · Privacy officers · Policy
Privacy regulations were designed around how humans collect, share, and use personal data — assumptions that break down when AI agents act autonomously, make decisions, and move data at speeds and scales no individual person could. Existing frameworks may lack the hooks needed to assign responsibility or enforce rights when no human is meaningfully in the loop.
Who should care: Lawyers · Compliance · General readers · AI governance · Policy · Privacy officers
The OECD's AI policy arm has published guidance aimed at establishing common frameworks for AI security, addressing threats such as prompt injection, model poisoning, and the risks posed by autonomous AI agents in deployment environments.
Who should care: AI governance · Lawyers · Administrators · General readers · Policy
The OECD is examining how to establish a shared conceptual framework for agentic AI systems — LLM-based agents capable of operating autonomously across physical and digital environments. As these systems grow more capable and attract broader investment, clearer definitional boundaries are seen as increasingly necessary.
Who should care: AI governance · Lawyers · Administrators · General readers · Policy