Why Government IT Architects Must Take a Bigger Role in Digital Transformation
Digital transformation of government is driving an enormous expansion in the collection and use of citizens' data -- which is why IT must step up and address key data-related issues.
- By Barry Devlin
- March 16, 2020
"Digital technologies are employed in the welfare state to surveil, target, harass and punish beneficiaries, especially the poorest and most vulnerable among them," according to a report published last October by Professor Philip Alston, UN special rapporteur on extreme poverty.
Before dismissing his report as applicable only to third-world countries, failed states, or those run by left- or right-wing despotic regimes, note that Alston's fact-finding trips since 2017 included visits to both the United States and the United Kingdom. The opening paragraph of the UN report singles out high- and middle-income countries as leading the charge in "digital governance" through technology-driven surveillance and control of welfare systems, the digitization of justice and immigration systems, online submission of tax returns and payments, and multiple other approaches, all based on ubiquitous data collection and digital transformation of governmental programs and processes.
Strategies designed to harass and punish beneficiaries are, of course, decided and defined by politicians and bureaucrats with the explicit or implicit support of the general public. However, the implementation of such strategies is a technology issue. Design and development decisions determine -- both intentionally and not -- how members of the public are affected. Those of us involved in delivering governmental IT systems, whether directly or via outsourcing, can play a role in how these systems operate and contribute to the public good.
This is just part of a bigger picture. Digital transformation makes data a core component of everything that is done. In the private sector, IT professionals (along with business users) are taking responsibility for data and analytics, but their IT counterparts in government often have no idea of the implications of technology or data decisions on the "business" (public service) or "users" (citizens). IT systems and data architects (among others) need to step up and assume a more central role in systems objectives and design.
To avoid "stumbling zombie-like into a digital welfare dystopia" as Alston warns, we must focus on two distinct areas -- data-related design and implementation issues.
Addressing Real User Needs
With interactions between government systems and the general public increasingly being driven online, the user interface becomes a critical component of design. We need to review assumptions about high-speed access to the internet and its affordability, the technological skills as well as the physical and mental abilities of end users, and the public's acceptance and trust in the system. IBM UK's CTO Dan Bailey and colleagues coined the term the digitally left-behind community to describe people who, for these reasons or others, have difficulty accessing digital services.
The UK's Universal Credit system, which is designed to consolidate multiple welfare schemes through a mainly internet-based system, has been condemned as unusable even by digitally savvy applicants. Some 46 percent of claimants reportedly need help with online applications.
However, beyond the user interface, IT must come to terms with a design premise of surveillance-like data collection. With a key political objective of eliminating welfare fraud, for example, claimants must continuously document and verify online their actions to ensure ongoing compliance. Failure to do so leads to suspended payments on the assumption of guilt.
Although such assumptions are political in nature, the systems implementers have completely failed to take account of the reality of life lived in poverty -- the time and money spent self-reporting is better spent putting bread on the table. Multiple deaths are already being linked to this toxic combination of systems objectives and design failures.
Avoiding Data Dystopia
A second failure of government IT architects of online citizen-facing systems is a limited understanding of the extent and quality of data available and in the poorly bounded use cases that are applied.
Assumptions about the data that can be collected, frequency of collection, and level of detail may be unreasonable in light of the technology (device features, network access cost, and speed, etc.) and skills limitations of the less fortunate who are the main data entry people.
Government IT professionals haven't grasped the "reasonable expectation" of law-abiding citizens about how their data may be used. We're not just talking about today's data. The analytics systems design and data availability of vast amounts of historical data has opened up a set of unintended consequences for the citizen beneficiaries of the systems. For example, consider how data usage boundaries can be transgressed by applying processing power to historical data to resurrect long-closed and/or forgotten social welfare cases.
In the United States, companies such as IBM, Oracle, and others have implemented new machine learning and artificial systems on behalf of state welfare agencies to combat current welfare fraud. However, they have also led to "zombie debt claims" for alleged overpayments in the 1980s being issued to thousands of poor seniors, often unable to either pay or contest the claims. This clawing back of funds is possible only because of the use of the old data in ways that could not have been even conceived a few years ago.
Of course, these are finally questions of ethics, law, and politics, but when IT systems (mis)use begins to directly impinge on citizens' life/death issues, especially in heretofore unexpected ways, systems designers and data architects need to step up to much broader responsibilities for the way the systems can be used.
Although privacy issues correctly receive much attention, especially in areas of justice and immigration, these aspects of data use and reuse require increased focus by governmental IT. The underpinning problem is that modern technology and big data stores have enabled bureaucratic decisions that had previously never been contemplated. In pursuit of the valid requirement of reducing fraud, technology providers have -- perhaps inadvertently -- unleashed the tools for a Kafkaesque, faceless bureaucracy.
Bridging the Business-IT Gap
These failures are similar to the long-standing business/IT gap in the private sector , a gap that is typically broader in governmental IT environments and significantly deepened by digital transformation. When information technology becomes core to the business, failures in communication between business and IT can easily spell disaster. Add the disparate types of user characteristics of government systems and it's clear how the challenges to delivering a viable and valuable government IT system may become insurmountable.
Addressing these challenges demands reworking our understanding of the requirements and architectural phases of systems design and, in particular, the data architecture of such systems, given the central role of data in digital transformation and the fact that citizen data is the foundation of governmental IT. The basic question is: how can the complex, disparate, and ever-changing uses of data in such systems be understood, documented, agreed, and managed?
Dutch consultant and author Martijn ten Napel contends that the only viable approach is to accept that "Data Architecture Is Really About People." This can be a difficult step for traditional IT designers and data modelers, whose basic instincts are to focus on methodologies and solutions.
As ten Napel explains, because when "data is the result of human activity -- reflecting all the quirks we humans have in our behaviors -- the answer is to create a framework where the value of using information is continuously evaluated against requirements derived from intended use." Rather than a traditional function-oriented approach, he describes a connected architecture, where connected emphasizes the necessary relationships between all the people involved in the design, use, and evolution of the data resource.
By applying connected architecture concepts to governmental systems, implementers could (and should) explicitly include the needs of the citizen users of these systems early in the design, allowing them to highlight and -- with political and bureaucratic support -- deliver systems better aligned to the common good of society.