How to Overcome Barriers to AI Adoption in Government

I was recently asked what is the biggest hurdle is to getting AI fully integrated into government systems. Here's my take by looking at this from multiple perspectives: Culture, Technology and Approach.

Culture Transformation

From a government perspective, I would say it's not a technology problem, but rather one of culture change and also of attracting AI talent into the agency. Third would be finding AI solutions that satisfy the agency's AI use cases while being able to integrate into the existing Enterprise Architecture and operational ecosystems.

For years the government workforce has been frustrated by the slow pace of change. Administration attempts to inject change via directives and Executive Orders has sparked discussion and planning activities with scattered vs. wholesale AI testing and adoption.

Government is risk adverse and rightfully so given the widescale damage mistakes can make. Agile methodologies have help shift IT thinking to deploy quicker and to forego doing Proof of Concepts (POCs) by moving right to delivering Minimally Viable Products (MVPs), where advanced functionality is deployed in smaller chunks that improve business processes, with enhanced features being released according to a schedule. This results in fewer "throw away" projects while gaining the same degree of knowledge, many by collaborating with industry experts to get hands-on learning.

Engaging the Next Generation

Many younger job seekers waved off considering jobs in the public sector -- creating a culture where most employees were pushing age 50 and rapidly moving toward leaving the system. Government knows it needs to attract and retain younger people with more modern skill sets to see the natural progression not just replacing staff but introducing new ideation and promoting the use of newer methods.

Technology

Solutions built upon open standards makes integrating new technology into an existing ecosystem relatively easy compared to the work it takes to convince people to change what they have been doing for many years. Attrition is also helping change attitudes, but knowledge is also walking out the doors. This further supports the need to move quickly.

Reuse is key to adoption in government! Any solution that aims to "rip and replace" will prove to be too expensive and lengthy to deploy. Government needs to change faster and deliver better results if AI is to show promised benefits.

Government has the "need for speed" as oversight entities (Congress, OMB, GAO, and the like) have little appetite for waiting decades - the government norm for many years. Delivering a positive Return On Investment (ROI) and demonstrating real business value by deploying quick hits is how change in government is happening today.

Taking a Horizontal Approach

Nothing I've pointed out is new nor surprising. The good news is AI offers government the opportunity to make wholesale change with minimal or at least controlled pain.

Government historically works in verticals -- this is because programs are funded by Department and are assigned to teams who are responsible for delivering on the objectives and requirements. This results in the inability to collaborate easily and forget about sharing data across programs... the government's data is locked-down in program silos.

The modern government needs to have a connected, contextual, and AI-ready data foundation for government-wide real-world view of connected data enriched with contextual information to support enterprise-wide data utilization and transformational decision making. The benefits of doing are enabling seamless access to connected and contextualized data across agency borders to enhance collaboration while modernizing data sharing and security controls.

I am calling on both government decision-makers and industry service providers to realize we are at a point in time where real change is coming... to paraphrase Eldridge Cleaver, "You are either part of the solution or you are part of the problem."