Generative Artificial Intelligence Standard
MnDOT Standard # IT-003
Effective: July 14, 2025
View/print signed standard (PDF)
Please go to the MnDOT Org Chart to find specific contact information: Org Chart.
Responsible Senior Officer: Technology and Data Leadership Committee Chair
Standard Owner: Chief Counsel
Contact: Deputy Chief Counsel – Technology and Data
Statement of standard
MnDOT is committed to aligning generative Artificial Intelligence (GenAI) use with the department’s core values, legal standards, policies and security practices. MnDOT employees and third parties working on behalf of MnDOT must use GenAI responsibly, ethically, and within the parameters outlined by this standard. Users are responsible for preserving the security of MnDOT’s data in accordance with MnDOT’s data management policies.
Data practices
Employees and third parties working on behalf of MnDOT must comply with the Minnesota Government Data Practices Act. It is imperative for employees using GenAI tools to understand the type of data they are working with before using the tool. Data that is private, confidential, protected, or otherwise not public must not be entered into any publicly available GenAI tool built on or continuing to train a Large Language Model. An employee or third party working on behalf of MnDOT that is unsure of the classification of the data they are working with must check the Business Data Catalog or contact the MnDOT Data Practices Team before using GenAI tools.
Employer-provided solutions
MnDOT employees may use specific, MnDOT-provided GenAI tools for data that is private, confidential, protected, or otherwise not public. The only approved tool for this purpose for is Microsoft Copilot Chat which shows the green shield with a checkmark in the upper right corner of the screen. Third parties working on behalf of MnDOT must work with MnDOT staff to ensure that any tools they use are approved by MnDOT before submitting private, confidential, protected, or otherwise not public data. Employees and third parties working on behalf of MnDOT are responsible for knowing which GenAI tool they are using before submitting data. Users must not use state email to register unsupported tools.
Ethics and bias mitigation
MnDOT employees and third parties working on behalf of MnDOT must practice responsible use of GenAI to avoid harm, discrimination, or violation of privacy. Users should use GenAI as a supplemental resource and not as a replacement for human judgment. All GenAI output used for decision making must be verified to ensure accuracy, relevance, and factuality and to mitigate risks such as AI hallucinations, bias, discrimination, misinformation, and negative impacts to people.
Ownership rights
MnDOT employees and third parties working on behalf of MnDOT may be waiving all rights of ownership to GenAI outputs that are created on behalf of MnDOT and used for services provided to the public. Users should consider acknowledging when GenAI has been used to create or modify content, either through blanket statements or integrated into the content.
Appropriate use
Use of GenAI must comply with this standard and the enterprise Appropriate Use of Electronic Communication and Information Technology Policy, HR/LR #1423.
Consequences
Employees who violate this standard or the Minnesota Government Data Practices Act are subject to discipline, up to and including discharge. Any person who willfully violates the provisions of the Minnesota Government Data Practices Act or any rules adopted under the chapter or whose conduct constitutes the knowing unauthorized acquisition of not public data is guilty of a misdemeanor 1.
Reason for standard
The purpose of this standard is to guide MnDOT employees and third parties working on behalf of MnDOT in the appropriate use of GenAI.
Applicability
This standard applies to all MnDOT employees and third parties working on behalf of MnDOT and using MnDOT data as part of their work for MnDOT.
Definitions
Artificial Intelligence (AI)
The simulation of human intelligence processes by computers. In AI, computers can “learn” to perform tasks that involve making judgments similar to those made by humans.
Applications of AI include pattern and image recognition, natural language processing (NLP), and business analytics.
Artificial Intelligence (AI) Hallucination
A response generated by artificial intelligence (AI) that contains false or misleading information presented as fact.
Generative Artificial Intelligence (GenAI)
A form of artificial intelligence that is designed to create new content or outputs by learning from existing data. GenAI models can produce original data, such as summaries, text, images, music, computer code, or other forms of creative work, based on patterns and information learned from large datasets.
Large Language Model (LLM)
A type of machine learning model designed for natural language processing tasks such as language generation. LLMs are language models with many parameters and are trained with self-supervised learning on a vast amount of text.
Natural Language Processing (NLP)
A machine learning discipline in which computers are algorithmically enabled to understand, process, and respond to human language in both spoken words and written text.
Not Public Data
Any government data classified by statute, federal law, or temporary classification as confidential, private, nonpublic, or protected nonpublic.
Roles and responsibilities
GenAI Users
- Understand and adhere to this standard and MNIT’s Public Artificial Intelligence Services Security Standard.
- Use only state approved or provided accounts on state approved or provided equipment for state work.
- Protect MnDOT data. Ensure that no private, confidential, protected, or otherwise not public data is entered into commercially available GenAI.
- Report the unauthorized use or disclosure of private, confidential, protected, or otherwise not public data to the Office of Chief Counsel Data Practices Unit.
- Ensure that use does not infringe on any copyright or intellectual property laws.
- Comply with software licenses, as applicable.
- Disclose where employees or members of the public are interacting with GenAI, the outcome and impact, if applicable, and the business purpose where GenAI is used.
- Ensure that GenAI use in decision-making can be explained and justified.
Standard Owner
- Identify key risks associated with GenAI usage, including data privacy breaches, intellectual property rights, and misinformation.
- Establish regular monitoring of GenAI tools for security compliance, data privacy, and adherence to ethical guidelines.
- Conduct periodic audits to assess AI's impact, mitigate risks, and review its compliance with policy guidelines.
- Create a governance structure to oversee GenAI tool usage, ensuring alignment with MnDOT goals and adherence to ethical standards.
- Ensure GenAI tools comply with applicable laws and policies, including data protection and privacy before approving their use at MnDOT.
- Review the standard every year, or sooner as necessary, to ensure the standard remains current.
- Ensure documents and training associated with the standard remain current.
- Communicate the requirements of the standard.
Resources
History and updates
Adopted
July 14, 2025 (Signed May 30, 2025)
Review of standard
The next scheduled review of this standard is due May 2026.