03/02/2026
The NDIS isn't just failing. It's succeeding at something else.
I’ve spent years watching a system that was built on the promise of agency and choice become something else entirely. It’s not broken. It’s optimising. It’s optimising for clean data over messy human reality, for plan compliance over life outcomes, for its own sustainability over our actual survival.
This is what I've come to call an Architected Intelligence. It mimics the form of human support—the plans, the reports, the line items—while gutting its substance. Your need for a friend becomes ‘social participation’ data. Your burnout becomes a ‘capacity building’ KPI. And now, with the government laser-focused on trimming its growth, the pressure is on to generate even more synthetic, cost-contained ‘success’.
It’s the ultimate feedback loop, and it’s actively harmful. As experts note, AI and data systems built for the "average" user systematically fail—and often exploit—people with disabilities, who are the original "edge cases". The NDIS has become a bureaucratic version of this: a system that classifies and processes but cannot truly personalise or care.
A.S. Social has done a deep dive into this called: The Synthetic Century. It connects the dots from the technical flaws of synthetic data and ‘model collapse’ to the real-world, legal cases where these algorithmic loops deny claims and push profit over people.
read the full article at:
When human touch became a biohazard, the solution was a mass migration to video grids.
