AI Civic Policy & Service Design
Service Design, Service Blueprint, Workshop Facilitation
TL;DR
The City of Austin needed AI policies before city staff could touch any GenAI tools. I advocated for a working session with executives to stress-test the policies. I ran it in 4 weeks, and the policies came out sharper: 3 added, 6 revised, 2 removed. Approved by the CTO. Turned into a staff user guide.
This working session was a part of a larger effort co-led by the City of Austin’s Office of Innovation and Technology Services agency.
There was a desire by City agencies and City staff to use Generative AI in 2023; however, the City was not yet equipped to implement and use AI tools in civic government.
City Problem
Before piloting AI tools could start, the City agencies responsible for technology needed to develop AI policies and guidance to support City staff and reduce inherent risk.
The pressure to put a good foot forward was high, and the unknowns were significant for the agency executives leading the rollout.
City Agency Problem
City staff often struggle to follow policies and guidance because they are difficult to read.
Not following City policy and guidance puts the City staff member at risk, as well as the City, its agencies, and residents.
City Staff Problem
Goal
(1) Develop and facilitate a design working session to help executives review, refine, and align on AI policies and guidance prior to piloting and submitting the policies for approval with the City Council. (2) In addition, create material that would make it easier for City staff to understand and comply with AI policies and guidance.
Team
Core team:
1 facilitator (me)
External stakeholders:
Agency executives
Tools
Service Blueprint, participatory design workshop, user guide
Project Duration
4 weeks
My responsibilities
Stakeholder interviews
Desk research
Service Blueprint
Workshop facilitation
Final report
Developing the Service Blueprint
To develop the service blueprint, I segmented City staff into 3 categories: using non-sensitive data, using sensitive data, and requesting a new AI tool.
For each user type, I mapped out the tasks they took from initial awareness to continual usage. After documenting the task, I mapped the draft policies and communication touch points in correlation to the task.
The map was the main artifiact used in the workshop.
Facilitating AI Policy Working Session
The goals for the working session were to: (1) gain understanding and clarity on the user experience, (2) find and document areas where compliance could be improved, and (3) identify the AI policies specific to users to create an AI policy user guide.
The session was built for two leaders tasked with submitting the AI policies to the City’s CTO for approval: Divya Rathanal, who led Digital Strategy & Solutions, and Daniel Cullota, who led the Office of Innovation.
Reviewing the AI service blueprint and policies with executives from Austin Technology Services and Office of Innovation.
Visual of the journey map with AI policies below, prompting questions and flagging areas of concern
Outcomes
3 new policies were added, 6 policies were revised, and 2 policies were removed. Three months after the working session, the policies were approved by the City’s CTO, and a user guide was developed to make it easier for City staff to comply with the new AI policies when leveraging AI in their work.
AI Policy
AI Policy User Guide
Learnings
I learned how civic policies can be tested and improved by overlaying policies with journey maps; the process had stakeholders reflecting on how people will experience the policies in real-life. Which was extremely helpful in aligning mult-agency leaders on areas for improvement within a short window of time.