Is interim the new permanent?

Jan 9, 2026

Sonali Shah, CEO of Cobalt and a leader in offensive security, discusses the future of AI in cybersecurity. She predicts that by 2026, AI will transition from a mere concept to a fundamental battleground. Shah highlights the importance of focusing on malicious intent rather than just data theft, warns about the risks of deepfakes in social engineering, and emphasizes the need for a mix of traditional and novel authentication methods. She also addresses AI's impact on talent pipelines and the necessity of retaining human oversight in complex security scenarios.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

AI Becomes Central Cyber Battlefield

Sonali Shah says 2026 moves AI from experiment to the central battlefield of cybersecurity.
Organizations face machine-speed threats that require autonomous defense and continuous offensive security.

INSIGHT

Old Problems, New AI Twists

Many AI security failures mirror old problems like excess privilege and key management.
New twists like indirect prompt injection arise when attackers poison external data AI ingests.

ADVICE

Block Intent, Not Just Code

Focus on blocking malicious intent, not just preventing data theft or code exploits.
Design defenses to detect manipulation of truth and harmful outputs from AI systems.

Get the Snipd Podcast app to discover more snips from this episode

Get the app