Microsoft Says New Cybersecurity AI Model Helps MDASH Score 95.95% at Half the CostThe Hacker News·Jul 28, 09:21 UTC · Jul 28, 2026Exploit / PoC45
OpenAI’s GPT-5.5 is out with expanded cybersecurity safeguardsHelp Net Security·Apr 24, 00:00 UTC · Apr 24, 2026Vulnerability30
OpenAI Unveils GPT-5.4-Cyber for Improving Cyber Defense With AIInfosecurity Magazine·Apr 15, 16:00 UTC · Apr 15, 2026Vulnerability30
Microsoft Launches Flurry of AI Security InitiativesInfosecurity Magazine·Jul 28, 12:45 UTC · Jul 28, 2026Vulnerability30
AI vulnerability discovery is pushing 2026 CVEs toward 66,000Help Net Security·Jun 19, 12:01 UTC · Jun 19, 2026Vulnerability30
New Attacks Trick OpenClaw AI Agent Into Running Code and Leaking SecretsThe Hacker News·Jun 11, 17:57 UTC · Jun 11, 2026Vulnerability30
Android malware detection collapses when the context stage comes outHelp Net Security·Jul 29, 00:00 UTC · Jul 29, 2026Malware30
GPT-Red beat human red teamers on a prompt injection testHelp Net Security·Jul 28, 12:10 UTC · Jul 28, 2026AI safety & security30
Prompt injection is becoming the XSS of the web agent eraHelp Net Security·Jul 17, 00:00 UTC · Jul 17, 2026AI safety & security30
OpenAI’s GPT-Red Automates Prompt Injection Testing to Harden GPTThe Hacker News·Jul 15, 00:00 UTC · Jul 15, 2026AI safety & security30
Indirect Prompt Injection in Web Content Targets AI AgentsInfosecurity Magazine·Jul 6, 15:00 UTC · Jul 6, 2026AI safety & security in the wild45
Frontier AI models collapse under multi-turn AI attacks, Cisco findsHelp Net Security·Jun 19, 11:19 UTC · Jun 19, 2026Vulnerability30