Opinion: The Government CIO Was Obsolete Before It Began

The Clinger-Cohen Act of 1996 was a response to the technology landscape of the 1990s. It centralized complex and expensive technology under a new role, the chief information officer. But 30 years later that role has become a dumping ground for anything tech related. We’ve also seen waves of new roles over the last decade from chief data officers (CDOs), digital officers, innovation officers and chief technology officers (CTOs) to the latest trend: AI officers. Some jurisdictions create distinct teams for these new officers, while others tuck them right under the CIO. Either approach adds complexity without fixing the underlying problem. It’s 2026 — we deserve roles built for this century.

WE NEED TO TALK ABOUT GOVERNMENT CIOs

The Clinger-Cohen Act celebrated its 30th birthday in February. It may have felt like a rational response to the technology landscape of the 1990s. But its attempt to corral complex and expensive technology investments along with 13 separate areas of responsibility into a single role was always ambitious.

As a result, it struggled from birth. Now, after three decades of legislative and executive patches, a vast accumulation of unmet Government Accountability Office findings, and the proliferation of related roles, it’s time to not just revisit but consider abandoning it.

COMPETING FOR THE CROWN OF GOVERNMENT INNOVATOR

The act was dated when it passed. Written with mainframes in mind, it landed as the web was shaking the foundation of our world and ushering in the first wave of e-commerce. By 2002, the E-Government Act of 2002 passed to address web-based services, and the security alarm bell of the Federal Information Security Management Act gave rise to chief information security officers.

The first federal CIO arrived in 2009 along with the cloud and then FedRAMP, followed by mobile. The CTO role shortly emerged as an executive innovation release valve as CIOs suffered under the sprawling burden of their role.

Next open data came to save us along with BIG data and Moneyball for government. CDOs emerged at the local level and formally joined the federal scene under the Foundations for Evidence-Based Policymaking Act in 2018. They were asked to focus on data governance instead of the analytics leadership that was actually needed.

Next, a customer-focused government would fix the problem with chief digital and/or experience officers, establishing yet another competitor for the crown of government innovator.

As these other roles were created to drive innovation and change, they were either tucked under the CIO shop or established elsewhere. If inside the CIO’s office, they were often unfunded and inherited whatever reputation and influence that shop had. If outside the CIO’s office, they often triggered turf wars and reputational spats. Neither approach resolved the structural issue.

Meanwhile, the average tenure of a federal CIO has hovered below two years (versus four-plus years in the private sector), while state CIOs are just over two years. The patchwork of legislative and executive order fixes have failed to address the fundamental failure of role definition, the original sin of 1996.

Now generative AI is upending the economics of technology with massive implications. And along with it comes yet another new role: the executively mandated chief AI officer.

OPERATIONAL OVERHEAD CRUSHES INNOVATION

When I was in San Francisco, we were on the cusp of recruiting our fourth CIO during my tenure as CDO. I made a suggestion: We should separate out the core IT functions and instead recruit for a director of IT services. I argued that we had put too much on the role, which implies innovation but its reality is operational. I lost that argument.

It’s more or less impossible for a CIO to simultaneously ensure operational excellence and innovate. One of our San Francisco CIOs emphasized the number of phones they managed. The CIO of one of my clients is very focused on the help desk. Another well-respected CIO struggles to fill basic positions and found themselves serving as admin on a key collaboration tool.

As long as the CIO is responsible for phones, Internet, conference rooms and other operational table stakes, they have a target on their back. In the worst cases, when the conference room doesn’t work, other leaders find it hard to trust that same leader with innovation.

At the state level, many CIOs are placed in the position of bottleneck and service provider. They control procurement approval and issue standards and policy but vary in scope and mandate. They are also caught in the operational trap of providing services to local governments.

Local jurisdictions face the same traps but with fewer resources.

RESTRUCTURING FOR THE GenAI ERA

While the role struggles to balance operations and innovation, the underlying assumptions in technology have radically changed. Since the arrival of cloud and software as a service, much of IT has commodified. And now GenAI is radically democratizing technical ability.

The original thesis of “centralize to save” is dissolving before our eyes with every new foundation model release. This is actually an opportunity: The business and program side of the house — the side that actually understands its needs — can now be positioned to be a responsible buyer of IT.

The GenAI era requires a restructuring of the CXO roles related to technology, data and innovation. Ideas on how to get there:

1. Separate table-stakes operational IT from everything else. This includes the basics that can be shared across departments (e.g., Active Directory, phones, Internet) and need to work all the time for everyone. This requires a steady, focused hand with boring standard service-level agreements. Combine it with HR and facilities functions. Recruit steady hands and don’t ask them to innovate.
2. Invert the authority model for technology. Empower program and mission staff to purchase technology with the right guardrails and abandon the role of CIO as sole purchase approver. This can include a preapproved vendor marketplace, standard easy-to-access contracts for service support and interoperability standards. I discuss more about this approach, including “managed options,” in “How government procurement creates tech stack chaos.”
3. Create a genuine center of excellence for major system migrations. Major migrations are the perennial source of billion-dollar government technology failures (and headlines), and they aren’t going away. But no single department or agency ever gets good at these, as by definition, they are rare. This work should be steered by an experienced, shared service function that can work across programs and even agencies building institutional learning and repeatable frameworks. This can’t be a glorified PMO. It must have real substance and experience in repeated technology migrations, along with hiring and procurement flexibility. Read more about how this could look in California.
4. Properly resource and define digital and data roles. The CDO and digital experience functions have real and distinct value, but only if they come with real authority and adequate staffing. Tucking them under an overburdened CIO or standing them up without a budget produces the same outcome: another title that can’t deliver.

This post is not a full solution. It is a call to reimagine the CXOs of government for the world of today, not the limits of the past. Let’s give Clinger-Cohen a birthday present: retirement.

Joy Bonaguro has spent her career working at the nexus of data, design, technology and policy. She now works as an independent consultant using every tool she needs to help make organizations work well. She served as a chief data officer at the city, county and state levels and has worked across public, private, research and nonprofit sectors. She has extensive experience working in complex, highly regulated organizations to accelerate use of data and technology in an agile, iterative, safe, ethical and durable manner. She’s an international speaker and writer on leading practices in data infrastructure, management, analytics, data science, privacy and responsible AI. Joy is a former member of Beeck Center for Social Impact + Innovation’s State CDO Network.