Artificial Intelligence at DOR

The Colorado Department of Revenue (DOR) is embracing Artificial Intelligence (AI) to enhance efficiency, improve services, and solve complex challenges for the benefit of our stakeholders and the public. Our AI Program is built on a foundation of responsible and ethical use, prioritizing data privacy, security, and human oversight.

AI systems are tools designed to augment human capabilities, learning from data to perform tasks that typically require human cognition. From generating content and assisting with data analysis to automating routine tasks and identifying fraudulent activities, AI offers immense potential to streamline our operations and provide valuable insights.

So far, 229 out of 1739 DOR employees have work accounts with Gemini. That's roughly 13%.

The Department of Revenue has assigned subject matter experts at each office to help with AI related inquires and tasks. They work to support the needs of their respective offices as well as the mission of DOR as a whole.

Here is the current list of AI Stewards across DOR:

Division

AI Stewards - Intranet

Guiding Principles.
Our guiding principles for AI use at DOR include:

Effective: Ensuring AI systems are reliable and deliver dependable outcomes.
Equitable: Striving for AI systems that support fair outcomes and manage bias.
Transparent: Communicating the purpose and use of AI systems proactively.
Accountable: Clearly defining roles and responsibilities for AI deployment and maintenance.
Human-Centric: Augmenting, not replacing, human judgment and critical thinking.
Privacy-Preserving: Protecting sensitive data and information.
Secure and Safe: Implementing best practices for data privacy and security.
Workforce Empowerment: Using AI to enhance employee capabilities and productivity.

The DOR AI Program is a dynamic and evolving initiative. We are dedicated to exploring innovative ways to use AI safely and responsibly, empowering our workforce, and delivering better services to the citizens of Colorado.

AI and You - Gemini

Our approved tool, Gemini, is here to work alongside you, fostering a culture of innovation and making our teamwork even more effective. You can learn more about what Gemini can do in our new feature that we’ll include in the All Hands called AI and You! During the AI and You Series, you’ll see best practices and innovative AI use cases across all DOR divisions. We'll share practical examples of how Gemini can help you, from drafting emails to summarizing documents, encouraging best practices across all divisions.

Approved AI Tools

Gemini is the sole approved AI tool for use by DOR employees. To safeguard DOR’s confidential, privileged, and sensitive data and information, employees are prohibited from using any free, public generative AI tools, such as ChatGPT, on DOR’s IT systems. All GenAI tool requests must undergo an OIT intake and risk assessment process before approval and use. Any current GenAI system or tool use that has not gone through OIT’s intake and risk assessment process must be discontinued until the system or tool meets policy requirements.

Google Gemini Implementation

As part of this exciting initiative, DOR is implementing Google Gemini for Workspace, a powerful AI tool designed to boost productivity by assisting with tasks such as drafting emails, summarizing documents, and brainstorming ideas. We are committed to providing comprehensive training and ongoing support to ensure all employees can effectively and responsibly leverage AI in their daily work

For DOR employees seeking more guidance, please reference the following resource folder: FAQs, Guides, and User Exercises

Below is a timeline of DOR’s implementation overall, but is subject to change by division:

CDOR Artificial Intelligence Policy (EDO-061)

This policy outlines the Colorado Department of Revenue’s requirements for the responsible, ethical, and secure use of Generative Artificial Intelligence (GenAI) systems and tools by all workers. It mandates OIT approval for all GenAI systems and strictly prohibits the use of public, unapproved GenAI tools on DOR IT systems. A key directive is the prohibition of using confidential or sensitive data (PII, CJI, FTI, PHI) in any AI prompt. The policy emphasizes human oversight in every aspect of AI usage. This includes human review, verification, and fact checking before any internal or external use of AI generated content.

_Acceptable Use, Risks, and Considerations of Artificial Intelligence .docx

Policy

This policy outlines the Colorado Department of Revenue’s requirements for the responsible, ethical, and secure use of Generative Artificial Intelligence (GenAI) systems and tools by all workers. It mandates OIT approval for all GenAI systems and strictly prohibits the use of public, unapproved GenAI tools on DOR IT systems. A key directive is the prohibition of using confidential or sensitive data (PII, CJI, FTI, PHI) in any AI prompt. The policy emphasizes human oversight in every aspect of AI usage. This includes human review, verification, and fact checking before any internal or external use of AI generated content.