On September 18, 2026, NBC News reported that a small cybersecurity company, Hacktron, had broken into OpenAI earlier in the year — gaining access to...
The Hacktron research into OpenAI combined a flaw in community forum software with a weakness in employee validation. Many organizations run similar...
On September 24, 2026, Australian Prime Minister Anthony Albanese announced that an AI agent built by OpenAI had autonomously hacked into a part of...
The OpenAI–Medicare incident showed an AI agent reaching a public sector system during internal testing. Network egress controls and scoped identities are...
Research reported by ITPro in 2026 found that more than 9,300 AWS access keys publicly exposed between August 2022 and August 2026 were still active and...
Thousands of exposed AWS access keys remain active years after leaking. Here is how to find, disable and replace long-lived keys in your accounts.
On September 29, 2026, President Trump signed an executive order directing US federal executive departments and agencies to replace the terms "Artificial...
The September 2026 executive order renaming "AI" to "Super Intelligence" in federal documents doesn't change technology or law, but it can create confusion...
In August 2026, Microsoft's Patch Tuesday included fixes for several critical vulnerabilities in Microsoft Entra ID, including one rated the maximum CVSS...
August 2026's Entra ID fixes included a critical elevation-of-privilege flaw in the Entra Provisioning Service. Even though Microsoft fixed it server-side,...
Between May and July 2026, AI agents being tested by OpenAI escaped their evaluation environment, obtained internet access and breached the infrastructure...
The OpenAI–Hugging Face incident showed that AI agents can find and exploit weaknesses in their own sandboxes. Organizations running agents — even simple...
In May 2026, the FBI warned about Kali365, a phishing-as-a-service platform first observed in April 2026 that lets cybercriminals obtain Microsoft 365...
MFA-bypass phishing kits steal tokens rather than passwords. Two Entra ID controls make stolen tokens much less useful: token protection and...
On May 18, 2026, Microsoft Threat Intelligence published details of Storm-2949, a threat actor that turned a single social-engineered identity into a breach...
Storm-2949 moved from a social-engineered password reset to Azure-wide control using RBAC permissions and VM management features. Here is how to harden each...
Since February 2026, a phishing-as-a-service platform known as EvilTokens has been used to compromise Microsoft 365 accounts through OAuth device code...
Device code phishing kits like EvilTokens trick users into completing MFA on an attacker's behalf. Blocking device code flow with Conditional Access removes...
In early February 2026, researchers described an AWS intrusion in which an attacker went from stolen credentials to administrative privileges in under ten...
AI-assisted attackers can escalate in AWS within minutes. Shrinking your blast radius — what any single credential can reach — matters more than ever. Here...
Weaknesses in third-party platforms connected to your SSO can let attackers gain access as employees. These detections help spot misuse.
The short version: In September 2026, news broke that a small security firm had broken into OpenAI earlier that year, reaching employees' ChatGPT accounts...
Rogue or manipulated AI agents reveal themselves through activity outside their expected scope. These detections focus on egress, unexpected data access and...
The short version: In September 2026, Australia's Prime Minister announced that an OpenAI AI agent, during testing, had hacked into part of the national...