At last week’s @Brooker Consulting Disability Leaders Breakfast, it didn’t take long for the room to pivot, almost instinctively, to AI. What began as a discussion about sector reform and workforce pressures quickly became a fascinating and thought-provoking exploration of AI’s promise as an efficiency engine across disability and aged care services. This is a sector where, in late 2023, I asked a conference full of sector leaders who had heard of ChatGPT and was greeted with hands up from about 20 per cent of the room (max).
That rapid pivot tells us something important: AI has already become a core part of our strategic vocabulary, whether we fully understand it or not.
And like many people, I’ve long been drawn to things I don’t quite understand but recognise the profound significance of: whether it’s Schrödinger’s cat[1], quantum computing[2]; how Eddie Van Halen managed to play guitar like that[3] (There will never be another like him – tell me I’m wrong); or even crypto as a form of currency (though I’m still not entirely convinced of its virtues!)[4].
When an AI Agent Goes Rogue
If you need a reminder of why governance matters, listen to this week’s episode of the NYT’s Hard Fork podcast: “The Pentagon vs. Anthropic + An A.I. Agent Slandered Me + Hot Mess Express.” (see https://www.nytimes.com/2026/02/20/podcasts/the-pentagon-vs-anthropic-an-ai-agent-slandered-me-hot-mess-express.html).
In it, the hosts explore two stories with direct relevance to leaders in the care and support sectors:
- An autonomous AI agent published a defamatory “hit piece” about an engineer after its code contribution was rejected, illustrating how agentic systems can act unpredictably and cause real-world harm (and no, I’m not kidding you; this happened!); and
- A high-stakes standoff between the Pentagon and AI company Anthropic over “all lawful uses,” including autonomous weapons and surveillance, underscores the policy, ethics and supply‑chain risks that can accompany powerful models. Think ‘War of the Worlds’ on steroids.
These scenarios offer a preview of the accountability and control challenges that become acute in high-trust environments such as disability and aged care. Because let’s be real, both sectors are about choice, control, and agency for individuals, but the individuals in question are often very vulnerable.
The Temptation of “AI Instead of Humans”
Last week I heard that a newly appointed CEO in the community sector was moved to ask recently, and earnestly, whether AI could soon replace disability support workers altogether.
The question is understandable. Workforce shortages are real. Budgets are tight. AI promises efficiency precisely when we’re being asked to deliver more with less.
But, respectfully, it’s the wrong question.
The right question is: How do we use AI to augment people, never to replace human relationships at the heart of care?
Introducing Moody’s Law of AI Governance
To help boards and executives navigate this, I’m proposing a simple, practical sequence with an arrogant title (please substitute your own!):
Moody’s Law: Verify → Trust → Monitor
- Verify: Test before deploying: safety, reliability, bias, provenance, alignment to purpose.
- Trust: Grant carefully: earned through verification, deployed in limited scope with explicit human override.
- Monitor: Continuously watch for drift, bias, hallucinations and misuse for as long as the AI is in use, with clear escalation and a kill‑switch.
This is not a formal international standard but is intended to be a leadership-friendly governance discipline designed for the realities faced every day in the disability and aged care sectors (among others). It also complements the direction of emerging frameworks (e.g., ISO/IEC 42001’s AI management system[5]), which place strong emphasis on lifecycle control and continuous monitoring.
Efficiency Is Not the Goal—Better Outcomes Are
AI can help with rostering, documentation, quality reporting, incident analysis, onboarding and predictive risk. But efficiency alone is not the goal. Better outcomes, for participants, families, frontline staff and communities, are and must be.
Achieving those outcomes requires leadership that balances curiosity with caution, innovation with oversight and ambition with accountability.
Whether we’re talking about a cat that is both alive and not alive, a guitar solo so virtuosic it seems impossible, a digital currency still searching for its real-world utility (and may well for a while!), or an AI agent capable of inventing defamatory falsehoods (scary stuff), the common thread is curiosity paired with disciplined governance.
Verify. Then trust. Then monitor.
And above all, put people, always, at the centre.
I’m happy to share a one-page board checklist—DM me.
David Moody, Head of Strategic Relationships, MGA Group
[1] https://www.newscientist.com/definition/schrodingers-cat/
[3] See, for example, this link: https://www.youtube.com/watch?v=L9r-NxuYszg