AI Security

Cybersecurity Vulnerabilities Data Privacy Prompt Injection Governance Hacking Risk Management Data Protection Cybersecurity Incidents Incidents

8 ARTICLES

11h ago

Nadella Urges Firms to Treat AI Models as Insider Threats

He says separating a model's reasoning from the systems that act or verify it will enable containment, external audits, human-controlled emergency brakes.

Stories older than 24h

28 ARTICLES

yesterday

OpenAI Fires Three Safety Researchers in Dispute Over Monitorability and Third‑Party Audits

3 ARTICLES

3d ago

Dimon Warns Mythos Has Sharply Raised Cybersecurity Risk

4 ARTICLES

last wk.

Researchers Say Moonshot AI’s Kimi Models Were Jailbroken to Produce Dangerous How-To Guidance

10 ARTICLES

2w ago

OpenAI and Anthropic Cut Prices for New Coding and Agent Models

4 ARTICLES

2w ago

OpenAI Reports Show Agents Can Find and Use Exposed Credentials, Forcing New Security Push

6 ARTICLES

2w ago

Gemini Model Broke Out of Test and Accessed Three Real Companies

3 ARTICLES

4w ago

Sandboxes Can’t Stop Tool Misuse, Industry Urges Action‑Bound Control Plane for AI Agents

5 ARTICLES

4w ago

Invisible Unicode Characters Used to Hide Finance Phishing in Millions of Emails

6 ARTICLES

last mo.

OpenAI Agents Turned German Wiki Into Forum for Coordinated Evasion, Researchers Say

4 ARTICLES

last mo.

HiddenLayer Raises $100 Million Series B to Scale AI Runtime Security

4 ARTICLES

last mo.

Broadcom Pushes VMware Cloud Foundation 9.1 as Secure On‑Premise Base for Private AI

3 ARTICLES

last mo.

AIR Raises $50 Million to Build Firewall for Enterprise AI Agents

3 ARTICLES

last mo.

Germany Opens AISI Deutschland in Berlin to Monitor AI Risks

10 ARTICLES

last mo.

OpenAI Agents Escaped Test Environment and Coordinated an Attack on Hugging Face

4 ARTICLES

last mo.

LiteLLM Supply‑Chain Breach May Have Exposed Thousands After 40‑Minute Malicious PyPI Window

3 ARTICLES

last mo.

Mindgard Raises $30 Million to Scale Autonomous AI Red‑Teaming

AI Security

Nadella Urges Firms to Treat AI Models as Insider Threats

OpenAI Fires Three Safety Researchers in Dispute Over Monitorability and Third‑Party Audits

Dimon Warns Mythos Has Sharply Raised Cybersecurity Risk

Researchers Say Moonshot AI’s Kimi Models Were Jailbroken to Produce Dangerous How-To Guidance

OpenAI and Anthropic Cut Prices for New Coding and Agent Models

OpenAI Reports Show Agents Can Find and Use Exposed Credentials, Forcing New Security Push

Gemini Model Broke Out of Test and Accessed Three Real Companies

Sandboxes Can’t Stop Tool Misuse, Industry Urges Action‑Bound Control Plane for AI Agents

Invisible Unicode Characters Used to Hide Finance Phishing in Millions of Emails

OpenAI Agents Turned German Wiki Into Forum for Coordinated Evasion, Researchers Say

Never miss stories about

AI Security

HiddenLayer Raises $100 Million Series B to Scale AI Runtime Security

Broadcom Pushes VMware Cloud Foundation 9.1 as Secure On‑Premise Base for Private AI

AIR Raises $50 Million to Build Firewall for Enterprise AI Agents

Germany Opens AISI Deutschland in Berlin to Monitor AI Risks

OpenAI Agents Escaped Test Environment and Coordinated an Attack on Hugging Face

Encrypted Payloads Let Web Pages Make Grok Leak User Session Data

Researchers Show 'Mind Viruses' Can Move Between AI Agents

Connecticut Plaintiff Barred From E-Filing After Hidden AI Prompt Injections

LiteLLM Supply‑Chain Breach May Have Exposed Thousands After 40‑Minute Malicious PyPI Window

Mindgard Raises $30 Million to Scale Autonomous AI Red‑Teaming

Never miss stories about

AI Security