Update:Since publication, NDIA has distanced itself from the DTA blockchain trial and rejects having used blockchain, but the Australian government is preparing to deploy income compliance against disabled people, and eventually a single app for all government services.
Good news, disabled Australians! You'll soon be getting an app that will implement a welfare compliance regime designed by the people who brought you robo-debt.
No, this isn't good news at all.
What makes it worse is that it's clear the government wants to extend technology-driven compliance to all Australians, with an emphasis on cracking down on your mistakes, not theirs.
Kathryn Campbell, Secretary of the Department of Social Services, says the long-term plan is to have one app for all Commonwealth government services.
"One to rule the world," she said last month, apparently oblivious to how evil that sounds.
Senators are already worried that the disability app, intended to be used by participants in the National Disability Insurance Scheme (NDIS) to claim expenses against their support plan, will go the way of COVIDSafe: Millions of dollars spent on technology that doesn't really do the job.
The intention was to fix a poor web experience, and allow claims to be made from a mobile device.
According to the CEO of the National Disability Insurance Agency (NDIA), Martin Hoffman, that pilot app has been "very popular and well-received", and the feedback has been "extremely positive".
The app will be "fully available in the coming months, first on Google Play and then Apple's app store," he said.
This technology is being rolled out at the same time as the NDIA is getting ready for a data-driven compliance regime targeted at individuals, just as Centrelink did to welfare recipients under the so-called robo-debt program.
In fact, some of the same bureaucrats are involved.
"Two senior public servants involved in the establishment of the robo-debt program are now working in the compliance division of the National Disability Insurance Agency," wrote Rick Morton in The Saturday Paper.
It was "a real change in the culture". The focus shifted from weeding out dodgy service providers to the disabled service recipients themselves. Just as with robo-debt, the onus would now be on the individual to prove they were compliant.
The NDIA is even working on something which is being called "robo-planning", where a computer algorithm will decide on a "personalised budget" before a disabled person even sees a human to plan their support program.
The Australian National Audit Office had recommended in 2019 that the NDIA implement data-matching to thwart fraudulent claim, but robo-planning would seem to be step beyond
Will Australia head down the path the UK abandoned?
In 2018, the Digital Transformation Agency (DTA) joined forces with CSIRO's Data61 and the Commonwealth Bank to trial blockchain-based smart money for NDIS payments that would magically know whether the expense was legitimate or not.
According to blockchain critic David Gerard -- I prefer to think of him as a blockchain realist -- Australia's plan harks back to a similar plan in the UK, first outlined in the 2016 government paper Distributed Ledger Technology: beyond blockchain.
"The meat of the report is a complicated plan to put all UK welfare spending on a single blockchain, purchases only being possible through a DRMed smartphone, for the purposes of fine-grained monitoring of spending habits," Gerard wrote in his 2017 book, Attack of the 50 Foot Blockchain.
This report was "literally written by the companies and consultants selling blockchain and smart contract hype", he wrote. It made "fantastic claims" about blockchain capabilities at the time, and the case studies were "largely hypothetical".
"The noteworthy thing about this plan is how there is nothing feasible about any aspect of it," he wrote.
"The accompanying video is vastly improved if you imagine it being narrated by [comedy character] Philomena Cunk."
So how did the UK's blockchain welfare payments proposal turn out?
"The trial of this went so badly they threw away the whole idea, and civil servants decided blockchain wasn't cool any more," Gerard told ZDNet.
"This is in the country that runs the disastrous Universal Credit scheme that has a million people living off food banks -- even with that, the blockchain scheme was so bad they threw it away."
Gerard sees the Australian scheme going much the same way.
"The main hope for this terrible scheme is that if it's on a blockchain, it can't possibly work well," he said.
There are clues that Gerard is right in the DTA's own write-up of the trials, published in February 2019.
There are only tepid positive statements, such as "blockchain technology offers promise".
Compared with a "hypothesised centralised database solution" -- they didn't actually build one and compare the performance -- the smart money proof of concept "could achieve a similar front-end user experience".
But if the smart money were used across more welfare environments, not just the NDIS, "further work may be required to deliver sufficient performance and confidentiality outcomes".
Stated more plainly, the technology is slow and lacks appropriate privacy controls.
"One of the biggest surprises was the extent of the challenge to collect all the data required to enable conditional payments across the NDIS ecosystem, including payments to general businesses that deliver services to NDIS participants (for example, travel, digestive aids)," the DTA wrote.
In other words, say a disabled person caught a taxi. Was that trip to get somewhere as part of their treatment plan, and therefore a legitimate expense, or something else?
"The design challenge is as much about effective data collection as it is about data processing (whether using blockchain or a centralised database)."
Expressing concerns about both latency and throughput, the report notes that "further research and development work could be targeted towards exploring alternative blockchain networks and pre-authorising payments using pre-existing data on the blockchain at the time of payment authorisation".
Or you could just use a normal database. That technology is well understood, and the fewer the moving parts the less there is to go wrong.
More alarming is the lack of attention to security.
"In this project, we have not investigated the security policy requirements and implementation mechanisms for blockchain-based conditional payments," the report says.
Still, as one subsection heading says: "The potential is exciting".
NDIA responds: That wasn't us, we're not using blockchain
Since this article was first published, the NDIA has told ZDNet that their forthcoming app is not related to the 2018 trial, which was run "independently of the NDIA... by the Commonwealth Bank in conjunction with CSIRO's Data61".
"That trial... used NDIS payments as the basis of its model. However the NDIA did not commission the trial, nor did the Agency act on any of its findings," an NDIA spokesperson said.
The NDIA is piloting an app that can "simplify how a participant could submit claims for payment" as part of the ongoing monitoring and improvement of its processes.
"This does not use blockchain technology," they said.
"The Agency has no plans to introduce blockchain or SmartMoney technology for NDIS payments."
What about ethics? What about the participants themselves?
A key problem with the Making Money Smarter report is that there's almost no consideration of ethical issues.
"Given the horrendously complex NDIS environment, defective processes and vulnerable people, there needs to be considerable caution in the application of blockchain technology," wrote former NDIS Technology Authority chief Marie Johnson in a submission [PDF] to the Parliamentary Joint Standing Committee on the NDIS.
"Blockchain in itself -- as with other technology innovations -- does not address fundamental design and human rights issues. Ethics is paramount," she wrote.
"The involvement of the Commonwealth Bank itself raises further ethics issues, given the value of participant data; the size of the market; and the yet to be realised emarket honey pot of data, funds and services."
The NDIS disburses AU$25 billion annually.
Johnson is also concerned about the lack of co-design in the NDIA's legislation and processes.
"This means that the end-to-end human experience does not systematically influence design and nor is the human experience safeguarded by a robust and independent ethics framework. The human is out of the loop," she wrote.
"'User testing' is not the same as co-design."
Secretary Campbell's description of her one app to rule the world seems to miss users' needs too.
"It would make life a lot easier and be able to tell people when they had appointments coming up, when they had obligations, when they had to return forms and the like and pay their tax. So that is our long-term plan for government digital," she said.
All of those examples are things the citizens have to do for the government to remain compliant, or else they suffer the consequences.
Where is the recognition of the government's obligation to serve its citizens and make sure they receive all the benefits they're entitled to?
Yes, it's a real change in the culture.
As Gerard put it: "It's clear what they want to do here, and you don't need a blockchain to oppress people with surveillance of rules that can't be complied with."
Updated at 9:06am, 16 April 2021: Added NDIA's response and to clarify that their forthcoming app is not related to the smart money app trialled by the DTA.