In a workshop with InnovateUS on January 28, Davyd Smith, IT Director for the State of Colorado Governor’s Office of Information Technology (OIT) and Jamie Kimes, an OIT contractor shared insights on the state’s successful pilot of Google's Gemini Advanced Generative AI (GenAI) tool across state agencies, along with their story on implementing responsible AI training for state workers with InnovateUS.
Why Google Gemini?
The decision to move forward with an enterprise AI solution came from a practical observation: state employees were already starting to use Generative AI tools. Rather than letting this happen ad hoc, the state chose to implement a statewide AI Policy and pilot Google's Gemini Advanced tool, largely because Colorado's government already worked with the Google Workspace.
“If we didn't come forth with a product, people are going to be using it anyway,” Smith said. “And there's danger in people actually using applications that are not part of your enterprise."
The pilot program
The program included 150 testers across 18 state agencies. The pilot's foundation rested on four key components:
-
A mandatory attestation agreement focusing on responsible AI use
-
Comprehensive training programs
-
Data tracking through regular surveys
-
Strategic communications support
Training and support
A key focus was ensuring that those engaging in the pilot were using AI responsibly. This meant engaging participants in comprehensive training. The training program was particularly robust, featuring:
-
Required completion of InnovateUS's Responsible AI for Public Professionals: Using Generative AI at Work course
-
Virtual learning sessions covering topics like fairness, bias, security, and privacy
-
Regular office hours that fostered cross-agency collaboration
-
On-demand training resources
The state worked directly with InnovateUS to tailor the training to their needs. Responsible AI for Public Professionals: Using Generative AI at Work is a two-hour, at-your-own-paced course designed for public sector professionals to learn how to integrate GenAI tools into their government work, enhancing public service delivery while managing risks responsibly.
"We worked with InnovateUS to actually customize it. It's branded inside our LMS. We have an introductory video from the Governor, we have the attestation that we talked about right in the video,” Smith said. “… It's a lot of fun... Really good training about the basic responsible use of AI.”
Pilot results
After analyzing over 2,000 use cases from the 90-day pilot, the results were compelling:
-
75% of users reported increased creativity
-
73% were able to focus on higher-priority work
-
70% felt more confident in their abilities
Approximately 20% of pilot participants identified as having accessibility needs. The AI tools provided remarkable benefits for these users, from improved communication capabilities to enhanced workplace confidence. For example, one visually impaired user expressed her excitement about the technology’s ability to write alternative text.
“There was an image of an ocean [and the learner] said, 'I almost felt like I could see it,'” Kimes said.
Moving forward
The success of the pilot has led to a structured statewide rollout. Colorado's approach to implementing Generative AI demonstrates how government organizations can responsibly adopt these powerful tools while prioritizing training, accessibility, and cross-department collaboration. Their pilot program not only improved workplace efficiency, but also fostered a more inclusive and confident workforce, providing a valuable model for other government entities considering similar initiatives.
The program's success shows that with proper planning, training, and support, government organizations can successfully implement AI tools while maintaining security, privacy, and responsible use standards.
