AI in the Enterprise:The Illusion of Control
Details
AI is moving into enterprise workflows quickly, while security teams are still working out what meaningful control looks like. Policies, approvals and guardrails can create confidence. But do they give organisations real visibility and agency over what AI systems do?
Join us for a facilitated fireside conversation on AI in the Enterprise: The Illusion of Control, exploring the question: How secure is secure enough? Two experienced practitioners will discuss what changes when organisations secure AI, who gets to define acceptable risk, and what it takes to trust AI with decisions that affect customers.
The conversation will be followed by a fishbowl, giving the audience a chance to join in and ask questions. Bring your practical questions and experience. No slides; just an open discussion across enterprise, academia and the non-profit sector.
## Guests
Mark Jones is Co-Founder and AI Deployed Engineer at MosaicalAI. After more than 25 years in cybersecurity, technology risk and resilience across financial services, government and critical infrastructure, he now helps organisations design, build and deploy AI agents that run securely and reliably. He weighs local models and hardware against frontier models to find the right balance of capability, resilience and cost.
Samm MacLeod is Group CISO at Iress, with more than 20 years’ experience in cybersecurity and technology risk across technology, financial services and critical infrastructure. She has held CISO and interim CISO roles at ME Bank, AGL Energy, Melbourne Water and Bulla Ice Cream. As VP Security (CISO) at Culture Amp, she helped guide responsible AI practices and led the company’s achievement of ISO/IEC 42001 certification. Her expertise includes cyber strategy, security transformation, AI governance and executive engagement, alongside advisory work with Netskope and cybersecurity industry boards.
## Questions to get us started
- Think of an AI tool you’d happily use at work today. What actually made you trust it?
- What’s different about securing AI compared with new technologies in the past?
- Who gets to decide what “secure enough” means, and are they the right person?
- What are enterprises getting right or wrong about AI today?
- What would you need to see before you let AI make a call about a customer without a human checking it?
Schedule: 5:45pm welcome; 6.00pm: pizza and drink and networking; 30-minute panel, followed by audience fishbowl.
Venue: Building 80, RMIT University, Melbourne.