22/10/2025
✊🏼✊🏼✊🏼
Last night we added the following addendum to our Notice of Concern lodged with the UNCRPD Committee in Geneva.
We appeal to them for intervention and swift safeguarding steps for the Disability community in Australia.
Machine learning, Ai and algorithmic decision-making in human services breach fundamental principles of due process, natural justice and human rights.
There is a time and place for these tools.
They do not belong in decisions that require human reasoning, empathy, judgement and complex case support. Machines cannot administer human justice and services.
Only humans can.
To Prime Minister Anthony Albanese, Mark Butler MP and Jenny McAllister we say only thing:
LOOK AT MOI PROIME MINISTA AND FRENS.
Look at MOI!
Anyway, these dibber dobbers are off to recover from yesterday.
Happy reading fam.
❤️🫡
Addendum to Notice of Concern: Use of Algorithmic Processes in NDIS Funding Decisions
Submitted to: United Nations Committee on the Rights of Persons with Disabilities
Submitted by: Sarah Langston, President, The Australian Neurodivergent Parents Association
Date: October 22, 2025.
Purpose of Addendum
This Addendum supplements the existing Notice of Concern lodged regarding systemic harm arising from the implementation of the National Disability Insurance Scheme (NDIS) reforms in Australia.
It raises a new and urgent concern regarding the possible use of algorithmic or automated decision-making processes in carrying forth the implementation of this legislative change; NDIS funding determinations, support needs assessments, and related administrative processes.
* Central Question for Immediate Clarification
Are algorithmic or automated processes being used to inform decisions at any stage of any NDIS process?
- This question must be directed to:
The Minister for the NDIS
The Minister for Health
The Prime Minister of Australia
The NDIA CEO and Deputy CEO
All relevant senior executives of the NDIA and Department
* Basis for Concern
There is a clear and reasonable basis to believe that algorithmic or automated systems may be influencing or determining participant funding outcomes.
This belief arises from:
• The unprecedented speed and scale of participant reassessments since September 2025
• The unusually high number of adverse decisions later overturned at the Administrative Appeals Tribunal
• The Australian Government’s significant investment in “AI-enabled service delivery” across Commonwealth programs
• The involvement of senior officials previously associated with the Robodebt scheme in drafting key NDIS legislative instruments
Together, these factors establish a reasonable and credible foundation for concern.
* Legal and Human Rights Implications
The introduction or use of algorithmic decision-making without transparency, consent, or oversight would violate the foundational principles of legality, procedural fairness, transparency, and individualised consideration under Australian administrative law, as well as key obligations under the United Nations Convention on the Rights of Persons with Disabilities (UNCRPD).
* Risk of Breach of the UNCRPD
The potential use of algorithmic decision-making within the NDIS gives rise to multiple, serious risks of breach of Australia’s obligations under the Convention.
- Article 3 General Principles
States Parties must ensure respect for inherent dignity, individual autonomy, and full and effective participation in society.
Automated or class-based decision-making negates autonomy and dignity by treating people as data points rather than individuals.
- Article 4 General Obligations
Governments must take all appropriate measures to modify or abolish discriminatory laws, customs, and practices.
The use of opaque algorithms risks embedding discriminatory practices within administrative systems.
- Article 5 Equality and Non-Discrimination
All persons are equal before and under the law and entitled to equal protection and benefit of the law.
Algorithmic systems that classify participants by statistical similarity rather than individual need create indirect discrimination and systemic inequality.
- Article 12 Equal Recognition Before the Law
Persons with disabilities have the right to recognition as persons before the law and to enjoy legal capacity on an equal basis with others.
Delegating decisions that affect rights and entitlements to an algorithm amounts to a denial of equal recognition and legal agency.
- Article 13 Access to Justice
States Parties must ensure effective access to justice, including through procedural accommodations.
Automated processes obscure reasoning, making it impossible for participants to understand or challenge decisions, thereby violating the right to a fair hearing.
- Article 19 Living Independently and Being Included in the Community
Automated funding reductions based on algorithmic classification may directly undermine the right to live independently and participate in the community.
- Article 28 Adequate Standard of Living and Social Protection
Algorithmic rationing of essential supports threatens participants’ access to housing, food, and basic supports, contravening Article 28.
* The Problem with Algorithmic Parameters
Algorithms cannot function without parameters. A human being must always first set parameters for an algorithm for it to function.
We tell an algorithm what to do, look for, research or organise. Prior to human ignition, it is static. A human being must first give it a set of boundaries to work within, in order to work at all.
Those parameters define the scope of the system and create classes by design.
Once parameters are set, the algorithm divides people into groups based on shared characteristics or patterns, and those classes shape outcomes.
Such classification directly contradicts the NDIS Act’s promise of individualised, reasonable, and necessary supports, as well as the UNCRPD’s requirement for individual dignity and equality before the law.
* Risk of Legal Liability and Class Action
If funding outcomes are being influenced or determined by automated systems, this could expose the Commonwealth to a future class action for unlawful or discriminatory administrative decision-making. This is because algorithms cannot function in the absence of the creation of a class.
It is reasonable to be concerned that this has already happened, with a key example being the cohort of 5-9 year olds with developmental Disabilities and/or who are Autistic issued with eligibility reassessments and with supports cut or reduced since the passing of the NDIS Bill.
Automated processes that generate or guide legal decisions without human judgment would likely breach the Administrative Decisions (Judicial Review) Act 1977, the Public Governance, Performance and Accountability Act 2013, and the NDIS Act 2013.
When coupled with the denial of procedural fairness, the risk of harm mirrors that seen in the Robodebt scheme, which was found to constitute unlawful debt recovery, breach human rights, and cause preventable deaths.
* Financial Benefit and Conflicts of Interest
The Committee should request disclosure of all entities involved in the procurement, licensing, or development of any algorithmic systems used by the NDIA or Department.
- Transparency is required regarding:
• Which companies or consultants are profiting from these systems
• The value and scope of related contracts
• Any financial, professional, or personal connections between these entities and government officials
The Robodebt Royal Commission demonstrated the danger of opacity and private profit in automated systems that make or influence public decisions.
* A Human System Must Remain Human
Australia’s common law system, for all its imperfections, is deeply and irrevocably human.
It is built on context, discretion, and at it's best, empathy - values that machines cannot replicate.
AI and algorithms may have limited uses. However, human matters must remain human. The NDIS Act is a human rights instrument, and the realisation of the promise of the UNCRPD. It must remain so.
Until full transparency, independent oversight, and participant consent are guaranteed, the only responsible and lawful course of action is to turn the robots off and to take a risk averse approach pending further information and community engagement.
* Requested Action by the Committee
That the United Nations Committee on the Rights of Persons with Disabilities:
1. Seek immediate clarification from the Government of Australia as to whether algorithmic or automated systems are used in any NDIS decision-making processes.
2. Request that any such systems be suspended pending independent review.
3. Recommend that Australia reaffirm and comply with its obligations under Articles 3, 4, 5, 12, 13, 19, and 28 of the Convention.
4. Urge Australia to ensure that all administrative decisions affecting persons with disabilities remain human-led, transparent, lawful, and rights-based.
Submitted respectfully,
Sarah Langston
President
The Australian Neurodivergent Parents Association
[Image description:
A meme made using a scene from the Australian TV show Kath & Kim.
In the image, two women stand in a kitchen. On the left, a woman with long brown hair (Kim) looks down sulkily. On the right, an older woman with short blonde hair (Kath) in a bright blue blouse leans toward her with a stern expression, hands on hips.
Text has been added in white boxes to mimic dialogue:
Kath (right): “now Kimmy look at moi please. we’ve talked about automating human services, haven’t we Kimmy? look at moi please Kimmy”
Kim (left): “but Muuuuum it’ll save me heaps of money and get me re-elected as Proime Minista”
The meme humorously re-imagines Kath & Kim as a political satire, referencing automation in government human services and the idea of a Prime Minister prioritising cost savings over ethics.]
