Freedom and guardrails for Citizen X

Hear from CIOs, CTOs, and other C-level and senior execs on data and AI strategies at the Future of Work Summit this January 12, 2022. Learn more

This article was contributed by Suzanne L. Taylor, vice president of Innovation and Emerging Technologies at Unisys

Many professions today are welcoming the rise of their Citizens X:  those ordinary citizens, curious or concerned about a particular discipline, who are empowered to collaborate with its credentialed professionals to make significant contributions. In technology, where the trend is sometimes called “the democratization of tech” or “no-code/low-code,” its benefits are great, but its risks must be recognized and mitigated.

I first noticed the rise of what I call “Citizen X” as an avid hiker, when I was reading the National Park Service site and came across an invitation to ordinary hikers to be “Citizen Scientists.” The park’s professional scientists welcomed us as amateurs to contribute to their work, suggesting how we could be helpful to them (take photos, record what plants were blooming and when, what wildlife, insects, or birds we sighted, and so on.)

Ever since then, I’ve observed the proliferation of Citizen X in many disciplines. There’s the Citizen Designer, with tools like Shutterfly and Vistaprint empowering even utter amateurs to create their own books or corporate logos and branding to be high quality and personalized.

There’s the Citizen Archeologist, invited to help archeologists find artifacts before they are lost to tides and time, making discoveries of untold riches and historical importance. And their underwater counterparts, Citizen Archeologists Underwater, a group of whom are thought to have discovered the oldest known wreck in Lake Erie.

Citizen Physicians are strictly for medical professionals, but who can deny the value of empowering ordinary citizens, also known as patients, to accelerate healthcare improvements? Patients can use their smartphones to assist in the diagnostic process and check their own vital signs and track their blood sugar levels. More advances are sure to come with the involvement of millions of patients collaborating with doctors, medical researchers, and pharmacists.

By and large, it is a positive trend. After all, a profession exists for the benefit of everyone. Doctors, lawyers, masons, manicurists – all are credentialed for the purpose of meeting the needs of the greater community. Among the people they serve are bound to be individuals whose ideas and work can further the profession. We know that innovation is often easier to achieve in smaller, nimble units and that innovation taking place at the grassroots can be remarkably creative and rapid – no small value in today’s need for speed.

Tech Risks

When it comes to technology, the Citizen X phenomenon poses risk as well as reward.

One of the most prominent examples is the Citizen Developer, which Gartner defines as “an employee who creates application capabilities for consumption by themselves or others, using tools that are not actively forbidden by IT or business units.” Not surprisingly, they find this freedom essential to their job satisfaction. They can automate processes that make their job easier, like completing their HR tasks, or faster. They are often in close proximity to users’ needs and know what innovation can bring value to users and the open-source tools needed to produce it. Not only that, but they are invaluable when it comes to prototyping.

But at the same time, they tend to be focused on a particular outcome and less attentive to matters that trained professional developers are scrupulous about. Such as: Is this software redundant with what others have already created? Can it be shared to be reusable by others doing similar work?  Is it created on a platform that is compatible with others? Is security embedded? Is it tested rigorously?

Boundaries are essential, given the risk that Citizen Developers may not be versed in and fully compliant with essential policies and protocols involving security, privacy, repeatability. IT supervision and formal governance are critical to any company sanctioning no code/low code developers. IT needs to create and rigorously monitor the underlying IT infrastructure.

Creating good governance for Citizen X’s work doesn’t require an entirely new invention – just a thoughtful adaptation of an organization’s current IT governance.  It should have the same level of protocol and documentation for everything:  development standards, roles and responsibilities, credentialing, compliance, and so on. If the Citizen Apps are intended for enterprise production or client use, IT needs to ensure that apps perform properly, meet high-quality standards, make proper use of corporate branding, and, especially critical in today’s environment, and meet the highest security standards. There may be looser governance for prototyping or limited use applications, but there still needs to be protection against compromising the enterprise.

With the trend towards a data-driven world, we see tools and platforms encouraging the Citizen Data Scientist, sometimes called Citizen AI Engineers. These platforms, tools, and services accelerate processes for using, analyzing, and deploying data. The leading cloud providers offer data science and AI cognitive services that enable not just professionals but also inexperienced practitioners to rapidly create machine learning and natural language applications.

This new technology is valuable given the perennial shortage of credentialed data scientists. But again, the low barrier to entry that these platforms provide also enables people to get in over their heads when they lack the expertise to fine-tune the data, understand the algorithms, and navigate pitfalls that skilled data scientists and AI engineers would be able to avoid.

The pitfalls range from algorithms that just do not work as intended when deployed on real-time data, driving up costs, to algorithms that could potentially cause harm. AI and ML are not deterministic systems. They cannot be thoroughly tested the way most other software can because they are designed to keep changing as they take in new data in real time. They cannot be tested for every permutation they are likely to encounter, and when they are out in the real world, they may not behave as predicted. They should be monitored for unexpected and unintended behaviors.

Ethical AI is a burgeoning discipline for good reason: At the extreme, there is potential for algorithms to cause real harm as they replace or augment humans in making critical decisions. Infamous examples include Amazon’s recruiting tool that was biased against women, a racially biased recidivism assessment tool, and Facebook’s discriminatory advertising algorithms.

Citizen Data Scientists are unlikely to be deeply schooled in the evolving discipline of ethical AI and can easily if unknowingly produce models that could endanger people, offend customers, or contravene norms and laws. More and more AI/ML regulations and recommendations continue to emerge, requiring companies to issue strict guidelines on the matter and requiring the oversight of trained data scientists and legal experts.

The trained data scientist, unlike the Citizen Data Scientist, embodies a sense of responsibility for the entire lifecycle of a model, guarding against faulty models resulting from, for example, overfitting or underfitting the data, and monitoring model drifts for subtle or not-so-subtle alterations.

Suzanne Taylor, PhD, is vice president of Innovation and Emerging Technologies at Unisys, where she focuses on applying emerging and disruptive technologies to business problems.

DataDecisionMakers

Welcome to the VentureBeat community!

DataDecisionMakers is where experts, including the technical people doing data work, can share data-related insights and innovation.

If you want to read about cutting-edge ideas and up-to-date information, best practices, and the future of data and data tech, join us at DataDecisionMakers.

You might even consider contributing an article of your own!

Read More From DataDecisionMakers

Source: Read Full Article