Cyberstats

Security Intelligence

When prompted to generate secure code, GPT-4o still produced insecure outputs vulnerable to 8 out of 10 issues.

Backslash Security

April 24, 2025

AI LLMs Vulnerabilities GPT-4o

Source

View Original Report

Published on 4/24/2025

Share or Copy this stat

Want More Statistics Like This?

Get the latest cybersecurity stats delivered to your inbox every week

Related Statistics

33% of respondents say AI agents shared sensitive or inappropriate data

37% of consumers worldwide identified the use of artificial intelligence in sophisticated scams, such as deepfakes, as their top concern in 2025.

Bitdefender•11/16/2025

AIScamsDeepfakes

64% of organizations report that AI and automation have increased their focus on the basics of cyber hygiene.

Swimlane•11/16/2025

AIAutomationCyber hygiene

Stay Ahead of Cyber Threats

Join 1,000+ security professionals getting weekly insights