Project purpose
This work aims to support the City of Portland to develop policies, best practices, training materials, communications, and public involvement, on responsible use of Artificial Intelligence and Automated Decision Systems. This work also includes developing initial priorities and gathering City stakeholders that need to be part of these discussions.
Project background
The 2019 City resolution that established the City of Portland’s privacy and information protection principles and 2023 surveillance technologies resolution define Automated Decision Systems (ADS) as the processes, set of rules, or tools based on automated processing of data to perform calculations, create new data, or to undertake complex reasoning tasks. This includes advanced methods like artificial intelligence and machine learning, visual perception, image generation, speech or facial recognition, and automated translation between languages.
City Council directed BPS’s Smart City PDX program to develop initial policies for the City’s use of Automated Decision Systems with the adoption of the Surveillance Policy via Resolution 37608 in February 2023, with the recognition that this work would be in partnership with BTS, OEHR and others.
An AI workgroup was created including city staff from the Bureau of Technology Services’ enterprise services group and the information security group, the Office of Equity and Human Rights, Office of Management and Finance’s risk management group, the Smart City PDX team, City Auditor’s Office, and City Attorney’s Office. City staff from the Bureau of Human Resources, and the City’s Administration Office representatives are invited as well.
What are the main concerns in the City?
This group initially identified general risks of using AI that could be used to start internal conversations. Some of them include:
- City liability for use, misuse, abuse of intellectual property (IP), copyright, internet-sourced content, images, source elements of data, sound, imagery.
- Data integrity, disinformation, and fabrication of artifacts in communications and false conclusions.
- Privacy and surveillance.
- Embedding unfairness in city process due to unchecked bias. For instance, in hiring practices, social programs, or city services.
- Breaking transparency and accountability laws and making the City liable to lawsuits.
What are the goals of this work?
- Respond to the Council directive in Resolution 37608.
- Develop citywide principles, guidelines, policies, FAQ for employees, leadership, and community.
- Create meaningful ways of public participation and digital education in equitable ways following the City of Portland’s code values and privacy principles.
- Contribute to the development of uniform use of terms and processes to address AI tools and other automated decision systems.
- Explore what uses of automated decision systems are highly sensitive or represent a public interest issue.
- Identify what is the most effective way to approach each bureau and support their work.
- Make sure that any new application does not compromise the City’s infrastructure or assets.
Project steps and timeline
2019 – Privacy and Information Protection Principles include language around Automated Decision Systems and AI.
2020 – Face recognition bans recognize intrinsic bias on artificial intelligence tools and place biometric data as sensitive information.
February 2023 – Surveillance technologies policy includes a directive to official work on Automated Decision Systems and Artificial Intelligence.
May 2024 – The City of Portland’s Automated Decision Systems work group kicks off activities.
Throughout 2024 – The Smart City PDX program is coordinating public events to provide input and comments on existing policies, best practices, and standards on use and procurement of Artificial Intelligence.