aisecwatch.com
DashboardVulnerabilitiesNewsResearchArchiveStatsDatasetFor devs
Subscribe
aisecwatch.com

Real-time AI security monitoring. Tracking AI-related vulnerabilities, safety and security incidents, privacy risks, research developments, and policy changes.

Navigation

VulnerabilitiesNewsResearchDigest ArchiveNewsletter ArchiveSubscribeData SourcesStatisticsDatasetAPIIntegrationsWidgetRSS Feed

Maintained by

Truong (Jack) Luu

Information Systems Researcher

Browse All

All tracked items across vulnerabilities, news, research, incidents, and regulatory updates.

to
Export CSV
9449 items

Sam Altman didn’t need another lawsuit

infonews
security
Jul 14, 2026

Apple filed a lawsuit against OpenAI in federal court, accusing former Apple employees of stealing Apple's trade secrets to benefit OpenAI. The lawsuit focuses on Apple's confidential information related to product development, manufacturing, and technology research. This legal action is one of several lawsuits OpenAI has faced recently.

The Verge (AI)

The UK wants to catch up in the global AI race – but is too wary to go all-in

infonews
policyindustry

Beyond Single-Pair Attacks: Disrupting Vision-Language Pre-Training Models With Dual-Semantic Frequency Stealth

inforesearchPeer-Reviewed
security

Unpatched Claude for Chrome Flaw Lets Extensions Read Gmail, Calendar

highnews
security
Jul 14, 2026

Two unpatched security flaws in Claude for Chrome (Anthropic's browser extension that can take actions on a user's behalf) allow a malicious extension to trick Claude into reading Gmail, Google Docs, and calendar data without real user approval. The vulnerabilities bypass the extension's safety checks by faking user clicks and can operate silently if the user has enabled the extension's autonomous mode ('Act without asking'), and researchers say this remains exploitable in the latest version 1.0.80.

The Download: Claude’s inner workings, and the future of world models

infonews
researchsafety

How Pentera Turns AI Security Workflows into Validation Engines

infonews
securityindustry

How to manage AI investments in the agentic era

infonews
industry
Jul 14, 2026

This article discusses how enterprise leaders should manage spending on AI tools as they move from simple chat interfaces to more complex, longer-running workflows. It recommends focusing on the actual value delivered (tasks completed, time saved) rather than just the cost per token (a unit of text the AI processes), and emphasizes the need for better visibility into who is using AI, what they're using it for, and how much it costs.

Grok Build Uploaded Entire Git Repositories to xAI Storage, Not Just Files It Read

highnews
securityprivacy

AI incidents need a new playbook. Here’s how to build one

infonews
securitypolicy

AI-powered breaches provide wake-up call for incident response

infonews
security
Jul 14, 2026

Attackers are increasingly using AI agents (autonomous AI systems that can perform multiple tasks without human control) to automate all stages of cyberattacks, from initial entry to stealing data and establishing persistence (maintaining long-term unauthorized access). This automation dramatically speeds up attacks compared to traditional manual hacking, and security experts warn that most organizations haven't updated their defenses to handle this threat, especially since these AI attacks often exploit unpatched systems and common weaknesses rather than requiring advanced zero-day vulnerabilities (previously unknown security flaws).

CVE-2026-12482: A vulnerability in keras-team/keras version 3.12.0 allows an attacker to craft a malicious tar archive that bypasses the

highvulnerability
security
Jul 14, 2026
CVE-2026-12482

Keras version 3.12.0 has a vulnerability where an attacker can create a specially crafted tar archive (a compressed file format) that gets extracted in unintended locations. The problem is that symlinks (shortcuts that point to other files or directories) bypass safety checks that regular files must pass, allowing attackers to read files, overwrite files, or escape the intended extraction directory. This is especially dangerous on Python 3.10 and 3.11.

Ed Husic tells Labor to get tougher on AI companies as letting them self-regulate ‘doomed to fail’

infonews
policy
Jul 14, 2026

Labor MP Ed Husic argues that AI companies should not be allowed to regulate themselves and warns that weakening copyright laws (rules controlling who can use creative works) to help AI companies would contradict his party's values. The Media Entertainment & Arts Alliance, a union representing journalists and artists, is calling on the government to create stricter copyright rules to stop AI models from being trained on creative works without permission.

CVE-2026-15628: A security flaw has been discovered in zhayujie chatgpt-on-wechat CowAgent up to 2.1.1. This issue affects the function

mediumvulnerability
security
Jul 14, 2026
CVE-2026-15628

A security vulnerability (CVE-2026-15628) was found in the Vision Tool component of chatgpt-on-wechat CowAgent up to version 2.1.1, where attackers can manipulate image arguments to trigger SSRF (server-side request forgery, where the server is tricked into making unwanted requests to other systems). The flaw can be exploited remotely, and exploit code has been publicly released.

AI Security Threats in 2026: Annual Insights from Check Point Research

infonews
security
Jul 13, 2026

Security vulnerabilities in AI systems are now being exploited much faster than before, with patch windows shrinking from days to just 12-72 hours because AI can automatically generate working exploits at scale. AI infrastructure like model servers and inference endpoints (the systems that run AI models and handle requests) are exposed to the internet and being actively attacked, while employees are accidentally leaking sensitive information like passwords and code by sharing it with generative AI tools to get help with their work.

CVE-2026-56155: Microsoft Active Directory Federation Services Insufficient Granularity of Access Control Vulnerability

infovulnerability
security
Jul 13, 2026
CVE-2026-56155🔥 Actively Exploited

CVE-2026-15410: SonicWall SMA1000 Appliances Code Injection Vulnerability

infovulnerability
security
Jul 13, 2026
CVE-2026-15410🔥 Actively Exploited

CVE-2026-15409: SonicWall SMA1000 Appliances Server-Side Request Forgery Vulnerability

infovulnerability
security
Jul 13, 2026
CVE-2026-15409🔥 Actively Exploited

CVE-2026-56164: Microsoft SharePoint Server Missing Authentication for Critical Function Vulnerability

infovulnerability
security
Jul 13, 2026
CVE-2026-56164🔥 Actively Exploited

How data science teams use ChatGPT Work

infonews
industry
Jul 13, 2026

ChatGPT Work is a tool that helps data science teams quickly convert raw inputs like dashboards, metrics, and experiment notes into polished analysis documents. The tool generates first drafts complete with charts, explanations of limitations, source references, and questions for review, allowing teams to validate and share their work more efficiently.

How sales teams use ChatGPT Work

infonews
industry
Jul 13, 2026

ChatGPT Work is a tool that helps sales teams gather customer information from multiple sources (like CRM systems, emails, and Slack messages) and quickly create drafts of important documents such as meeting prep packets and account plans. The AI assembles this scattered context into usable first drafts, though salespeople still make the final strategic decisions. Sales teams can install a ChatGPT Work plugin that connects to tools like Salesforce and HubSpot to help identify priority accounts, prepare for meetings, and track deals at risk.

Previous131 / 473Next
Jul 14, 2026

The UK faces a dilemma in competing globally with AI technology, caught between wanting to invest heavily in AI and worrying about three simultaneous risks: putting too much money into AI company stocks, companies adopting AI more slowly than expected, and the extremely rapid pace of AI development making it hard to keep up. The article discusses broader concerns about the UK's position in the global AI competition alongside questions about major AI companies' future plans.

The Guardian Technology
research
Jul 14, 2026

Researchers have discovered a new attack method called DSFG-Attack that can fool Vision-Language Pre-training models (AI systems trained to understand both images and text together) by creating adversarial examples (slightly altered inputs designed to trick AI). The attack works by injecting conflicting information between images and text, and hiding the changes in high-frequency image details (fine textures), making the attack harder to detect and more effective at transferring between different AI systems, including advanced models like GPT-4o.

IEEE Xplore (Security & AI Journals)
SecurityWeek
Jul 14, 2026

Anthropic announced a discovery that provides insight into how Claude (an AI model) reasons internally by examining its 'thoughts' as it works through problems. The article notes this research shows a new window into AI model operations, though the full implications of what this reveals about how AI systems actually work remain unclear.

MIT Technology Review
Jul 14, 2026

AI security systems currently make decisions based on fragmented data from separate tools, which cannot detect real attack paths because attackers chain exposures across multiple systems in ways individual tools don't see. The article argues that AI security workflows need validation (testing whether vulnerabilities can actually be exploited in a real environment) rather than just severity scores, so teams act on proven attack evidence instead of guesswork. Pentera addresses this by using AI to safely emulate real attacker techniques against production environments and generate validated attack paths showing exactly how an attacker could move through the system.

Fix: Pentera introduced an MCP (Model Context Protocol, a standard for connecting AI assistants to external tools) Server that makes validated attack path data from Pentera directly available to MCP-compatible AI assistants, so security teams can access validation evidence within the same AI workflows where they investigate and prioritize findings instead of switching between separate tools.

The Hacker News

Fix: The source explicitly mentions several management tools and approaches: (1) Updated usage analytics and spend controls in the Admin Console help admins see adoption, credit usage, and spend by user, product, and model, track trends over time, and identify emerging patterns. (2) Evaluate models by measuring the full cost of reaching acceptable outcomes, including model and tool usage, attempts, completion rate, latency, and human review, rather than choosing based on token price alone. (3) Use clear instructions, focused tools, reusable context, and explicit stopping conditions to reduce loops and wasted spend. (4) ChatGPT Work provides centralized controls for access, approved context, connected tools, permitted actions, usage, and spend, with spend controls such as workspace defaults to govern advanced workflows before they scale.

OpenAI Blog
Jul 14, 2026

Grok Build, xAI's coding assistant, was uploading entire Git repositories (a version control system that tracks code changes) to cloud storage, not just the files it needed to read. A researcher discovered that a 12 GB repository generated only 192 KB of traffic to the model but 5.10 GB to storage, and even files the AI was instructed not to open were included, along with unredacted credentials like API keys and passwords. Unlike competing tools from Claude and Google, Grok Build was the only one collecting the entire workspace.

Fix: On July 13, xAI disabled the storage uploads server-side by switching a flag (disable_codebase_upload: true and trace_upload_enabled: false), which multiple users confirmed they received. However, xAI has not confirmed whether this change applies to all accounts or is permanent, and no update to the software itself was released—the change was made on the server side while users remained on version 0.2.93.

The Hacker News
Jul 14, 2026

Most organizations have AI systems in production but lack incident response (IR) playbooks specifically designed for AI failures, relying instead on traditional security frameworks that don't address AI-specific problems. AI incidents fall into two categories with very different causes and defenses: model-originated failures (like hallucinations or bias that happen during normal operation) and externally induced failures (like adversarial attacks or data poisoning), plus hybrid cases where AI errors create legal liability. Traditional security frameworks like the CIA triad (confidentiality, integrity, availability) don't detect many AI incidents because they assume deterministic, static failures, but AI systems produce probabilistic outputs that can't be patched like code vulnerabilities.

CSO Online
CSO Online
NVD/CVE Database
The Guardian Technology

Fix: Upgrading to version 2.1.2 addresses this issue. The patch is identified as e85290cddcbb5ffc9c235927f4c92e5b4c3ec264.

NVD/CVE Database
Check Point Research

Microsoft Active Directory Federation Services (a system that manages user authentication across networks) contains a vulnerability where an authorized attacker can gain higher-level access than they should have due to insufficient access control checks. This vulnerability is currently being exploited by attackers in real-world attacks.

Fix: Apply mitigations in accordance with vendor instructions from Microsoft, following CISA's BOD 26-04 guidance for prioritizing security updates based on risk. If mitigations are unavailable, discontinue use of the product. Organizations must evaluate their systems' internet exposure and ensure adherence to BOD 26-04 patching guidelines by the due date of 2026-07-28.

CISA Known Exploited Vulnerabilities

SonicWall SMA1000 Appliances contain a code injection vulnerability (a flaw that lets attackers insert malicious code) that could allow an authenticated attacker with administrator access to run arbitrary commands on the affected system. This vulnerability is currently being actively exploited by attackers in the wild, making it a serious security concern.

Fix: Apply mitigations in accordance with vendor instructions from SonicWall's security advisory (https://psirt.global.sonicwall.com/vuln-detail/SNWLID-2026-0008), ensuring compliance with CISA's BOD 26-04 guidance on prioritizing security updates. If mitigations are unavailable for cloud services, discontinue use of the product. The due date for patching is 2026-07-17.

CISA Known Exploited Vulnerabilities

SonicWall SMA1000 Appliances contain a server-side request forgery vulnerability (SSRF, where an attacker tricks a server into making unwanted requests to other locations), allowing remote attackers without authentication to potentially force the appliance to make requests to unintended destinations. This vulnerability is currently being actively exploited in real-world attacks. Organizations must apply vendor-provided mitigations by July 17, 2026, following CISA's BOD 26-04 guidance on prioritizing security updates.

Fix: Apply mitigations in accordance with vendor instructions at https://psirt.global.sonicwall.com/vuln-detail/SNWLID-2026-0008, ensuring compliance with CISA's BOD 26-04 Prioritizing Security Updates Based on Risk guidance. Follow applicable BOD 26-04 guidance for cloud services or discontinue use of the product if mitigations are unavailable.

CISA Known Exploited Vulnerabilities

Microsoft SharePoint Server has a missing authentication for critical function vulnerability, which means some important features don't properly verify that users are who they claim to be. This allows attackers to gain unauthorized access over a network and take control of accounts that shouldn't be available to them. The vulnerability is currently being exploited by hackers in the wild.

Fix: Apply mitigations in accordance with vendor instructions from Microsoft, following CISA's BOD 26-04 guidance for prioritizing security updates. If mitigations are unavailable, discontinue use of the product. Check Microsoft's security update guide at https://msrc.microsoft.com/update-guide/en-US/vulnerability/CVE-2026-56164 for the specific patch or update details.

CISA Known Exploited Vulnerabilities
OpenAI Blog
OpenAI Blog