GIW Federal 2025: AI Risks — Safeguarding Services and National Security with James Kay

As a key leader within the ASD, James Kay's work extends beyond traditional signals intelligence, focusing on defensive cybersecurity measures, standards, and the responsible adoption of AI.

“ASD has historically been a collector of signals intelligence, but we’re building a new defensive capability—one focused on standards, technical guidance, and safeguarding Australia’s digital ecosystem.”

The AI Landscape: Opportunities and Risks

James highlighted that AI adoption offers enormous potential for innovation across government and industry—but also introduces significant security challenges.

“Government and industry were incredibly aligned on the potential of AI—but also on the extraordinary opportunities it creates for those who seek to do us harm if we don’t implement it correctly.”

The fireside chat reflected on lessons from Singapore Cyber Week, where more than 90 nations gathered to discuss AI and cybersecurity. James praised the regional alignment on safe AI adoption, emphasizing the principles of secure-by-design and secure-by-default:

“As we implement AI models, we need to make sure we’re not opening a back door for a threat actor.”

Domestically, James highlighted the recent National AI Plan:

“It’s positive to see we’re embracing this technology for its extraordinary potential—but we must make sure cybersecurity is brought along for the ride.”

James offered actionable insights for ensuring AI is implemented safely, ethically, and inclusively:

• Embed cybersecurity alongside AI deployment: Vulnerable populations will inevitably interact with AI-enabled tools.

• Apply ISM guidance: While designed for government systems, these controls can benefit private organisations depending on size and complexity.

• Vendor accountability: Ask vendors if they follow ASD guidance and the Information Security Manual.

• Sign up for alerts: cyber.gov.au provides early warnings, patching advice, and threat intelligence.

• Partnership program: Organisations with an ABN can join for early access to security advisories.

“If you’re not already signed up for alerts at cyber.gov.au—do it now. And if you’re an organisation, sign up for our partnership program. It gives you early warning on threats we can’t publish publicly.”

Governance and Oversight

James stressed that robust governance ensures AI is used responsibly:

“The Inspector-General of Intelligence and Security can come into my office any time and review everything we do. That oversight, along with the Parliamentary Joint Committee, ensures we use AI ethically and responsibly.”

Internal policies complement external oversight, helping ensure AI deployment is both secure and accountable.