Skip to content
All Expert guides & datasheets

The AI assurance gaps nobody is talking about

Andrew Robinson’s CyberCon 2026 presentation

 

CyberCon Melbourne 2026 · Wednesday 14 October 2026

 

AI didn’t break governance. It exposed its weaknesses.

In this CyberCon Melbourne 2026 presentation, 6clicks Co-Founder Andrew Robinson challenges nine assumptions holding organisations back, from treating governance as control to relying on annual audits. Explore practical ways to close the gaps, strengthen accountability, and realise more value from AI.

 

By submitting this form, I agree to 6click’s Privacy Policy.
Overview

Nine gaps. How many are you closing?

Policies, risk registers, and annual audits can create confidence without proving AI performs as intended. Strong governance balances risk and value.

Andrew Robinson examines nine gaps in governance, risk, compliance, accountability, liability, security, assurance, transformation, and society, and how to address them.

shield-check_bg-secondary-500_fill-secondary-950_card

Governance is about control

shield-check_bg-secondary-500_fill-secondary-950_card

Risk works the same way it always did

shield-check_bg-secondary-500_fill-secondary-950_card

Compliance is merely a necessary evil

shield-check_bg-secondary-500_fill-secondary-950_card

Fire someone. Problem solved.

shield-check_bg-secondary-500_fill-secondary-950_card

The vendor carries the liability

shield-check_bg-secondary-500_fill-secondary-950_card

The security team has us covered

shield-check_bg-secondary-500_fill-secondary-950_card

But we audit it every year

shield-check_bg-secondary-500_fill-secondary-950_card

Incremental improvement is enough

shield-check_bg-secondary-500_fill-secondary-950_card

Societal implications don’t apply to us

 

 

 

CONFERENCE SPEAKER

Meet Andrew Robinson

 
6clicks co-founder and former CISO, Andrew Robinson,
brings deep expertise in cyber security, governance, risk, and compliance.
 
At CyberCon Melbourne 2026, he presented on the AI assurance gaps receiving too little attention,
with a focus on the independent assurance of deployed models and autonomous agents.