AI Security

GPT-6 Astra Cuts Hallucinations, Prompt Injection Risk

September 04, 2026 1 min read 4 views
cybersecurity news newsapi
OpenAI's GPT-6 Astra reduces factual errors to 4.2%, but hidden prompt injections still succeed in about 1 of 12 scenarios, posing ongoing security risks.

Was this article helpful?

0 found this helpful

Related Articles

AI Slop Isn't Just for Social Media; It's Ruining Cyberattacks, Too

1 min read

Why AI Confidence Is Not Validation

1 min read

Measuring Human Cryptanalysis Effort

1 min read

Ready to protect your organization?

Get started with Phishlens and run your first phishing simulation today.

Schedule a Demo