How AI is impacting cybersecurity. Does it?
summary
Research shows GitHub Copilot (Codex) often suggests insecure code: experiments across Python and C using CWEs and CodeQL find 44% vulnerable top answers, including a persistent SQL injection example and insecure choices like MD5. Model behavior traces to being trained on public GitHub repos, which also enables privacy leaks, dependency/version drift, data‑poisoning and model‑stealing risks. There is little logging or accountability for when AI‑generated snippets enter production, so static analysis and secure review remain mandatory.
AI-written summary · about
chapters (11)
- 0:00AI and cybersecurity impact
- 1:47AI hype and ChatGPT adoption
- 8:15Cybercriminals leveraging AI for malicious code
- 13:36AI in enterprise security
- 22:50AI risks and ethics
- 27:08GitHub Copilot and Codex deep dive
- 31:29AI code vulnerabilities
- 41:10Developer risks and accountability
- 48:44Cybersecurity risks and opportunities
- 53:07AI security predictions
- 1:00:12Closing and Q&A
related talks
53mHow AI is impacting cybersecurity. Does it?
48mAI Impact on Cybersecurity and Privacy
58mSecuring Code in the Age of AI
41mCode Security Reinvented Navigating the era of AI
38mCode Security Reinvented: Navigating the era of AI
30mHuman vs AI: How to Ship Secure Code
26mSmarter Code, Sneakier Risks: Supply Chain Security in the Age of AI
10mAI Coding & Cybersecurity — Why AI coding tools aren’t ready for production yet