City of Moreno Valley Technology Services
AP # 7.10
AI (ARTIFICIAL INTELLIGENCE) USAGE
ADMINISTRATIVE PROCEDURE
Approved by: City Manager
November 19, 2024
PURPOSE: To establish guidelines for City Employees for the use AI (Artificial Intelligence) while performing responsibilities associated with the City of Moreno Valley.
I. DEFINITIONS AND TERMS
AI (Artificial Intelligence) : Any algorithm for developing statistical , behavioral, logic- based, large language, neural network, or other models that use computers to perform tasks without explicit instructions. AI often relies on patterns and inferences to perform work , predict better outcomes, and mimic the operations of a human brain.
Algorithm: A set of steps, such as mathematical operations or logical rules placed in sequence to accomplish a task.
City Employee: Any person or entity compe nsated by the City or volunteering with the City. This includes, but is not limited to, full-time employees, part-time employees, interns, temporary employees, elected officials, volunteers, commissioners, contract employees, partners, vendors, contractors, and other affiliates.
Machine Learning: A type of AI in which computers use data to “learn” tasks through algorithms. The learning process is also called training.
II. INTRODUCTION
Generative Artificial Intelligence (AI) is a new branch of AI technology that can analyze, identify patterns, answer questions, make recommendations, summarize findings, and generate content - such as stories, poetry, images, voice, music , etc. - at the request of a user. The City recognizes the general availability of AI tools and the opportunity for a responsible approach to their use that acknowledges the efficiency benefits they offer while minimizing the risks around AI bias, inaccuracy, privacy, and cybersecurity.
AI is a new and powerful technology that promises to multiply your efforts and handle rote tasks for you so that you can focus on higher -level tasks; it can also spark innovation and give you different perspectives to consider. Likewise, AI is fraught with inaccuracies and is already being used to spread disinformation, malware, and to manipulate people through fake writings, voices, and pictures.
AI presents staff with an opportunity to work better, faster, and smarter. However, because the technology and the laws surrounding it are evolving and present unknown risks, its adoption comes with ethical considerations. By keeping the guidance below in mind when using AI tools, we can ensure the safe and responsible use of AI by all employees.
City of Moreno Valley Technology Services
AP # 7.10
AI (ARTIFICIAL INTELLIGENCE) USAGE
ADMINISTRATIVE PROCEDURE
Approved by: City Manager
November 19, 2024
As we all learn to use AI responsibly, please be an active participant in suggesting helpful guidelines to refine the ones below.
III. GUIDANCE ON THE USE OF AI
The items below are guidelines for the use of AI that every employee is expected to know and follow as they utilize AI in their wor k. Please consult your supervisor or manager if you have questions about utilizing AI.
A. Remember that all information you enter into an AI tool is subject to a Public Records Act (PRA) request. The information you enter into ChatGPT (Generative Pre-trained Transformer) and other AI systems is saved in your account , can be used by the system , and may be viewed by others. Therefore, it is considered “released to the public” for purposes of the PRA. Presume that anything you submit to an AI tool could end up on the front page of a newspaper.
Only provide information to AI tools that is already approved for public disclosure.
This includes any text, photos, videos, or voice recordings. Be mindful that the AI output may include unexpected personal information from another user, so be sure to remove all personal information before publishing.
B. Review, revise, and fact-check via multiple sources any output from AI. The human user is responsible for any material created with AI support. The use and consequences of using AI is a human responsibility and cannot be defe rred to a system. Each City Employee is responsible for their use of AI and AI work products.
The City maintains trust with its residents and partners by providing accurate information. Review and fact-check all outputs you receive from AI. Users should consult trustworthy sources to confirm that the facts and details in the AI-generated content are accurate. Trustworthy sources include official City documents and peer- reviewed journals.
C. Appropriately cite your use of AI. The City shall be clear and transparent when it uses AI. This can often include citing that you used AI in creating a product. There are several tools available to check if documents were created by AI, they allow residents to check if City documents were generated by AI regardless of whether staff cites its usage or not. To build trust with residents, staff needs to be proactive in communicating its usage of AI. Residents finding out on their own that a document was created using AI can cause reputation harm to the City.
1. Staff must cite the AI when a substantial portion or significant point of the content used in the final version comes from AI.
2. Any statements used as fact must cite a credible source rather than AI.
City of Moreno Valley Technology Services
AP # 7.10
AI (ARTIFICIAL INTELLIGENCE) USAGE
ADMINISTRATIVE PROCEDURE
Approved by: City Manager
November 19, 2024
Credible sources include official City documents and peer -reviewed journals. Consult your supervisor for other trustworthy sources (e.g., newspapers, blogs, or datasets).
3. All images and videos must cite any AI used in their creation, even if the images are substantially edited after generation . Citations for images and video must be embedded into every frame of the image or video. For support on how to do this, consult the Media Division.
4. AI can be cited as a footnote, endnote, header, or footer. Citations for text- generated content must include:
i. Name of the AI system used (e.g., GPT-3, ChatGPT-4, Google Bard, Stable Diffusion , Dall -E, Bing Chat, GitHub CoPilot, Lensa AI , etc.); and
ii. Confirmation that the information was fact -checked. For example:
“This document was drafted with support from ChatGPT. The content was edited and fact-checked by City staff. Sources for facts and figures are provided.” or “Some of this content was drafted using ChatGPT. Al l facts, figures , pictures , and statements were reviewed by the creator to be accurate.”
5. For further information, see U.S. Copyright Office for Works Containing Material Generated by Artificial Intelligence (88 FR 16190).
D. AI system responses are based on patterns and relationships learned from large datasets derived from existing human knowledge, which may contain errors and is historically biased across race, sex, gender identity, ability, and many other factors.
City users of AI need to be mindful tha t AI may make assumptions based on past stereotypes and needs to be corrected.
E. Create an account only for City use. Because usage is subject to public records requests, City Employees shall have an account for all AI usage in their role at the City. This account should not be used for any personal work. Employees can use their City email address for City usage, or they can create a shared account using a different work email address. F or example, Technology Services may create a shared account using [email protected] as the email address. Regardless of whether a shared or work email address is used to create an account, employees must use a unique password for the service. Like any other account that uses a City email address, the password should not be the same password used to log into any City device. Data breaches have already occurred on ChatGPT (March 2023).
F. Data for AI usage is typically stored on the cloud, so the device used to access the City-specific account is not relevant. If staff use personal devices or accounts to conduct City work, the records generated are still public records subject to search and disclosure whether they are stored locally on the device or in a cloud account associated with the device/account.
G. Staff should be open to responsibly incorporating AI into their work where it can City of Moreno Valley Technology Services
AP # 7.10
AI (ARTIFICIAL INTELLIGENCE) USAGE
ADMINISTRATIVE PROCEDURE
Approved by: City Manager
November 19, 2024
make services better, more just, or more efficient. For example, a tool like ChatGPT can help staff go from an outline to a draft City Council memo quickly, enabling staff to focus more time on the analyses and findings that inform recommendations to the City Council.
IV. ASSESSING RISK IN AI USE CASES
A. AI risk will be determined by at least three key factors:
1. Risk of information breach: The potential harm if the information exchanged with a n AI system is released to an unintended audience. This can include entering personally identifiable information, sensitive records, or confidential business information. Any information entered into AI is considered “released to the public” for purposes of the Public Records Act and waive s any applicable exemption. If you would not share the information in a public forum, don’t share it with AI.
2. Risk of adverse impact: The potential harm of using the output for a decision, task, or service. This impact can be different for different populations and should be considered from an equity lens, such as adverse impacts on people of a certain race, age, gender identity, or disability status.
Not only can AI be biased, but it can also provide false information. In general, if AI is used with City processes that can alter an individual or community’s rights, freedoms, or access to services, it should be thoroughly reviewed by multi ple users before any document is finalized or action is taken.
3. Integrity of the information: The results from using AI may be incomplete, wrong, misleading , unreliable, inaccurate , or reduce the importance of human empathy, discretion, or judgement. All an alysis, recommendations, and results from AI shall be reviewed by the user and their judgement applied to the integrity of the work product.
B. Assigning risk based on the above key factors:
1. If the information exchanged with a n AI system would be harmful to a person or community if made public, it is a high or intolerable risk. Services like ChatGPT have been compromised in the past and leaked personal information. Until private applications with higher security are deployed in the City, all information exchanged with AI has a reasonable risk of being compromised.
2. Mid-risk information includes non -identifying and non -confidential information. For example, a simple email response or instructive documents often contain only general information that would not present any risk if made public. Mid-risk impact includes tasks associated with drafting internal messages, internal documentation, and idea generation. These tasks City of Moreno Valley Technology Services
AP # 7.10
AI (ARTIFICIAL INTELLIGENCE) USAGE
ADMINISTRATIVE PROCEDURE
Approved by: City Manager
November 19, 2024
can be sped up with the support of AI but require many more steps before reaching a public impact.
3. High-risk information includes personally identifiable information (e.g., full name, birth date, email address) and confidential business information that may have larger implications for City processes. Until a private application is deployed with security measures approved by the Technology Services Division , no high -risk information shall be provided to a n AI system. High-risk impact includes tasks associated with official City documents or messaging. Presentations generated by AI are at least high - risk because they go beyond text to include images. The high-risk label also includes uses that require substantial editing and review before usage. These tasks require thorough review at the time of generation before us e in any work context. Special care should be taken when a task may impact individuals differently across factors such as race, age, gender identity and disability.
4. Prohibited risk information includes highly sensitive and identifying information. This includes data such as credit card numbers, bank account information, social security numbers, and other information that requires rigorous security measures and compliance standards b efore being processed. Prohibited risk impact includes tasks that undermine trust in the City through false statements or news; deny people due process such as in resource allocation, job evaluations, and purchasing decisions; or expose the City to substan tial security or legal risks. AI does not have reasoning behind the content it produces and cannot justify a decision; these activities fall to the human user of AI who is responsible for its use.
5. If you are using AI with City processes that can alter an i ndividual or community’s rights, freedoms, or access to services, it is at least high risk and should be thoroughly reviewed before any document is finalized or action is taken. Additionally, any action that could reasonably lead to the City engaging in legal infringements on intellectual property is prohibited.
V. POSSIBLE AI USE CASES
A. Currently, there are tasks where AI can be of invaluable help and some tasks where it should not be consulted for assistance. Below are tasks that AI is well -suited to assist with. Understanding what AI is good at will help you determine possible cases where AI can increase your productivity and accuracy. Please keep these examples in mind as you determine if AI is the correct tool to assist or not.
B. Quickly applying known rules to separate queries where answers or actions are known versus queries needing human decisions. Known case: First Responders and Dispatch Centers use AI to handle non-emergency calls for a 10-12% reduction in City of Moreno Valley Technology Services
AP # 7.10
AI (ARTIFICIAL INTELLIGENCE) USAGE
ADMINISTRATIVE PROCEDURE
Approved by: City Manager
November 19, 2024
calls for dispatchers.
C. Perform work where the methods are well -known. Known case: Medicaid must recertify tens of millions of members quickly. AI can reach out via several channels (e.g., phone, text, TTY, email, etc.) to track where people can actually be contacted for recertification.
D. Quickly correlating multiple input sources to make routine decisions. Known case:
Vehicle sensors in a shipyard track the size, number, routes, frequencies, schedules, weights, etc. inside the shipyard to ensure optimized traffic flow.
E. Generating draft documents for specific purposes from known data or facts. Known cases: Provide ChatGPT with a topic and request the outline of an informational training session on that topic. Provide a Generative AI system with a table of metrics and ask for a written narrative based on analysis of the metrics.
F. Sifting thru large amounts of data for statistics, relationships, and correlations, then make a recommendation based on the correlations or data.
G. Making predictions based on machine learning training. Note that the source and recency of training data is of immense concern when these systems are used.
Known case: Given a portrait, request what the person will look like after 20 years of disappointment.
H. Learn repetitive basic human tasks and perform them repeatedly and flawlessly, thereby r educing human errors (omission and commission) . Known case:
Automatically check multiple systems to determine if a vendor owes the City money before they are allowed to bid on a new project.
I. Performing analysis or tasks with greater precision than humans can. Sensors and robots can measure and act in microscopic detail . Also, large amounts of data can be transcribed from audio recordings and summarized in major themes or content.
J. Performing known tasks consistently for long periods. Known case: Computer security logs are scanned in real -time, 24 hours a day , for indications of cybersecurity breaches.
K. Recognizing patterns in data through the application of standard analysis techniques.
L. Translate English-language questions into machine-language queries. Known case:
Census GPT accepts English questions and provides answers to complicated census data questions.
M. Performing mundane tasks. Known case: Monitoring rooms or park s for noise levels to capture patron usage data.
VI. PROHIBITED USES OF AI
A. AI shall not be given information or data that is not already in the public domain .
This includes written, pictorial, audio, and digital data. No private, controlled, or confidential data shall be added to a publicly accessible AI ser vice or training City of Moreno Valley Technology Services
AP # 7.10
AI (ARTIFICIAL INTELLIGENCE) USAGE
ADMINISTRATIVE PROCEDURE
Approved by: City Manager
November 19, 2024
model.
B. AI output shall not be used to impersonate individuals or organizations without their written permission.
C. Using software code generated through generative AI shall not be used in production until fully reviewed and tested for proper functionality and security.
Any such use shall be properly and thoroughly documented.
VII. AI TRAINING RESOURCES
Occasionally, training in AI or specific AI tools will be offered. If the timing is not convenient for your project, work with your supervisor to obtain access to an AI training course.
Khan Academy provides high -quality, free online training. Access khanacademy.org and search for “generative AI”.
For developing prompt engineering skills —the ability to provide good input for great AI output—the following site provides free, high-quality training: learnprompting.org.
In November 2023, the GovAI Coalition was established to give local government a voice in shaping the future of AI and ensure that AI development is geared toward the benefit of society. The coalition involves the collaboration of pubic and private entities. The City of Moreno Valley is a member of the GovAI Coalition.
The GovAI Coalition is composed of over 600 public servants from over 250 local, county, and state governments that represent over 150 million Americans across the nation united in our mission to promote responsible and purposeful AI in the public sector. For resources like AI vendor reviews, AI government use cases , and AI vendor agreements see the “Deliverables” heading under:
https://www.sanjoseca.gov/your-government/departments-offices/information- technology/ai-reviews-algorithm-register/govai-coalition?_t_ip=52.167.144.211