United States - Washington - Responsible AI Policy

Overview

The Seattle Responsible AI Policy, formally known as the Artificial Intelligence (AI) Policy, establishes a comprehensive framework for the ethical and effective use of artificial intelligence solutions within the City of Seattle's operations and service delivery. This policy, which was updated and announced as part of the broader 2025-2026 AI Plan, aims to define processes that enable AI systems to address urban challenges and enhance municipal services, while simultaneously upholding the City's core values, particularly those outlined in its AI Guiding Principles. It provides clear, actionable guidance for all City staff, including full-time and part-time employees, interns, consultants, vendors, contractors, partners, and volunteers who provide City services or act on behalf of the City. The overarching goal is to ensure that AI adoption is responsible, secure, and transparent, thereby fostering public trust and confidence in City services. This policy represents a significant evolution from earlier guidance, including an interim policy from April 2023 and a specific Generative AI Policy released in November 2023, integrating lessons learned from numerous AI pilot projects already underway across various departments.

The policy's scope is intentionally broad, covering all forms of AI solutions used by the City, not just generative AI. It emphasizes that while AI offers immense potential for innovation and efficiency, it also carries inherent risks, such as the potential to perpetuate inequity and bias, which could erode public trust. To mitigate these risks, the policy mandates adherence to established City security, privacy, and responsible use practices from the initial stages of procurement through deployment. It also seeks to align AI governance with existing City data governance, security, privacy, and contracts programs, ensuring a cohesive and integrated approach to technology management. The policy is maintained under the authority of the Chief Technology Officer (CTO) ordinance (SMC 3.23.030), underscoring its legal foundation and the commitment of the City to a structured and accountable implementation of AI technologies.

Definitions

The Seattle Responsible AI Policy incorporates several key definitions to ensure clarity and consistent understanding across all City operations. "Generative Artificial Intelligence (Generative AI)" is defined as a class of computer software and systems, or functionality within systems, that leverage large language models, algorithms, deep-learning, and machine learning models. These systems are characterized by their ability to generate new content, including but not limited to text, images, video, and audio, based on patterns and structures derived from input data. This definition also extends to systems capable of ingesting input and translating it into another form, such as text-to-code systems. While the policy's principles apply broadly to AI technologies, specific policy statements may focus on generative AI systems.

Another crucial term is "Responsible AI Use," which the policy describes as an ethical approach to developing, procuring, and utilizing AI solutions. This approach is centered on community needs and considers a multi-faceted set of values, including equity, innovation, efficiency, transparency, privacy, security, and resiliency, all in the context of delivering exceptional City services. The policy also implicitly defines "AI Solutions" as the various systems employing artificial intelligence that are acquired, configured, developed, operated, or maintained to provide municipal services or address City challenges. Furthermore, the concept of "Human-in-the-Loop (HITL)" is fundamental, requiring that any outputs generated by AI systems undergo human review before being used in an official City capacity. The "Racial Equity Toolkit (RET)" is also a defined procedural requirement, mandating its application before using generative AI tools to assess potential impacts and ensure outputs are accurate and free from discrimination or bias. These definitions collectively establish the conceptual bedrock for the policy's principles and implementation guidelines.

Governance and Institutional Framework

The governance and institutional framework for the Seattle Responsible AI Policy are designed to ensure robust oversight and accountability for AI use within the City. The Chief Technology Officer (CTO) is explicitly designated as the primary authority responsible for the creation, maintenance, and compliance with this policy, operating under the mandate of the CTO ordinance, SMC 3.23.030, and in coordination with the Mayor's Office. This centralized authority ensures a consistent approach to AI governance across all City departments. The CTO or their designee holds the power to inspect AI solution usage to verify adherence to responsible use principles and policy requirements, with the ability to mandate alterations or cessation of a partner's AI activities if necessary to protect the City and its residents.

To further strengthen this framework, the City of Seattle is establishing a new AI leadership position and a cross-departmental AI Governance Group. This dedicated leadership and collaborative body will play a critical role in overseeing the implementation of the policy, fostering inter-departmental cooperation, and ensuring that AI initiatives align with the City's strategic priorities. The Seattle Information Technology Department (Seattle IT) also plays a vital role, being responsible for providing essential training and skill-building resources related to AI foundations, approaches, and solution development. This commitment to training ensures that City employees are adequately prepared to responsibly create and manage AI-driven municipal solutions. The policy also mandates the alignment of AI governance with the City's existing data governance, security, privacy, and contracts programs, creating a cohesive and integrated regulatory environment for all technological deployments.

Key Focus Areas

The Seattle Responsible AI Policy is built upon a foundation of seven core guiding principles that represent the City's values regarding how AI solutions are acquired, configured, developed, operated, or maintained. These principles are: Innovation and Sustainability, emphasizing the responsible exploration of AI to improve services and advance beneficial outcomes; Transparency and Accountability, ensuring that AI system development, use, and deployment are compliant with laws and publicly documented; Validity and Reliability, committing to AI systems that perform consistently and accurately with ongoing evaluation; Bias and Harm Reduction and Fairness, acknowledging AI's potential for inequity and mandating evaluation through an equity lens; Privacy Enhancing, prioritizing the protection of personal data throughout the AI lifecycle; Explainability and Interpretability, striving for AI models and outputs that are easily understood and communicated; and Security and Resiliency, ensuring AI systems maintain data confidentiality, integrity, and availability through robust protection mechanisms. These principles serve as the ethical and operational compass for all AI-related activities within the City.

In addition to these guiding principles, the policy explicitly defines prohibited uses of AI to safeguard against potential societal harms. Certain AI applications are entirely forbidden for City operations due to concerns about their performance, potential biases, and infringement on fundamental rights. These prohibited uses include, but are not limited to, emotion analysis, social scoring systems, and integration with autonomous weapons systems. Furthermore, the policy bans the use of AI in hiring processes and facial recognition products, reflecting a proactive stance against technologies deemed to carry unacceptable risks for discrimination and privacy violations. By clearly delineating both guiding principles and prohibited applications, the Seattle Responsible AI Policy aims to strike a crucial balance between harnessing AI's transformative potential and upholding the City's commitment to equity, privacy, and public trust.

Implementation Framework

The implementation framework for the Seattle Responsible AI Policy is designed to embed responsible AI practices into the daily operations and procurement processes of the City. A core requirement is that all AI solutions must be acquired through the City's approved procurement channels, which now include specific AI-related considerations to ensure consistency with the City's standards and principles. This centralized approach to acquisition helps to vet technologies for compliance before they are integrated into City services. A critical procedural safeguard is the "Human-in-the-Loop" (HITL) requirement, which mandates that all outputs generated by AI systems must undergo review by a human prior to each use in an official City capacity. Departments are also required to document their HITL review processes, demonstrating how these reviews adhere to the policy's principles. This ensures human oversight and accountability for AI-generated content.

Furthermore, the policy emphasizes the application of a Racial Equity Toolkit (RET) prior to the use of any generative AI tool. This toolkit is designed to facilitate an evaluation of potential impacts, ensuring that AI-generated content is accurate, fair, and free of discrimination or bias against protected classes, aligning with the City's broader Race and Social Justice commitments. The Seattle Information Technology Department (Seattle IT) is tasked with providing essential training and upskilling resources to City employees, covering AI fundamentals, various AI approaches, and specific solution building. This commitment to workforce development ensures that staff are equipped with the knowledge and skills necessary for responsible AI creation and management. The policy also stresses the importance of aligning AI governance with existing City data governance, security, privacy, and contracts programs, promoting a holistic and integrated approach to technology management and compliance across all municipal functions.

Monitoring and Evaluation

The Seattle Responsible AI Policy establishes a robust framework for the ongoing monitoring and evaluation of AI systems to ensure their effectiveness, adherence to principles, and continuous improvement. A key component of this framework involves the establishment of specific metrics to evaluate AI pilot projects. These metrics encompass a range of factors including performance, accuracy, reliability, adoption rates, user satisfaction, scalability, return on investment (ROI), and alignment with overarching business goals. By systematically measuring these aspects, the City can assess the tangible benefits and potential drawbacks of AI initiatives, allowing for informed decision-making regarding their broader implementation or modification. This data-driven approach ensures that AI deployments are not only innovative but also deliver demonstrable value and meet desired outcomes.

Beyond pilot projects, the policy mandates ongoing evaluation of system accuracy throughout the entire development and deployment lifecycle of AI solutions. This continuous assessment is critical for managing, governing, and auditing the reliability of AI systems, ensuring they perform consistently and as expected under various conditions. The Chief Technology Officer (CTO) or their designee is empowered to inspect the usage of AI solutions across City departments. This oversight function is crucial for verifying the City's responsible use of AI and ensuring strict adherence to the policy's stipulations, thereby protecting both the City and its residents. This comprehensive monitoring and evaluation strategy underscores Seattle's commitment to adaptive governance, allowing for early detection of issues, fostering iterative improvements, and maintaining public trust in the City's use of artificial intelligence.

Penalties, Liability, and Appeals

The Seattle Responsible AI Policy outlines clear consequences for non-compliance, emphasizing accountability for all individuals and entities operating within its scope. For City staff, vendors, contractors, partners, and volunteers who fail to adhere to the policy's requirements, a range of disciplinary actions may be imposed. These can include restriction of access to AI systems or data, formal disciplinary measures up to and including termination of employment for City employees, or termination of contracts for vendors and partners. The Chief Technology Officer (CTO) is responsible for enforcing compliance, often in coordination with individual division directors and department leaders, ensuring that accountability is shared across the organizational structure. This tiered approach to enforcement allows for responses proportionate to the severity and nature of the non-compliance, from minor infringements to significant breaches of policy.

While the policy explicitly details penalties for non-compliance, it is important to note that as a municipal policy, it primarily focuses on administrative and contractual consequences within the City's operational sphere. It does not directly establish broader legal liability frameworks or specific appeal processes akin to those found in statutory law. However, the policy's emphasis on transparency, accountability, and adherence to existing City laws and regulations implies that any actions taken under the policy would be subject to standard administrative review processes and potentially existing contractual dispute resolution mechanisms for vendors. The overarching goal of these enforcement provisions is to ensure that the City's commitment to responsible AI use is upheld, mitigating risks and protecting public trust by ensuring that all AI-related activities align with Seattle's values and legal obligations.

Relationship to Other Instruments

The Seattle Responsible AI Policy is not an isolated document but is strategically integrated within a broader ecosystem of City governance and legal instruments. It explicitly states its alignment with the City's ongoing data governance, security, privacy, and contracts programs. This integration ensures that AI solutions are not developed or deployed in a vacuum, but rather adhere to established standards for data handling, cybersecurity, personal information protection, and procurement ethics that apply to all City technologies. This cohesive approach prevents fragmentation in regulatory oversight and promotes a unified standard of responsible technology use across all municipal functions.

Furthermore, the policy operates under the direct authority granted by the Chief Technology Officer (CTO) ordinance, specifically SMC 3.23.030, which establishes the CTO's role and responsibilities. This statutory backing provides the policy with a strong legal foundation, underscoring its binding nature within the City's administrative structure. The current Responsible AI Policy also represents an evolution of earlier City initiatives, building upon lessons learned from an interim AI guidance issued in April 2023 and, more notably, subsuming and updating the City's Generative Artificial Intelligence Policy, which was initially released in November 2023. This iterative development reflects Seattle's adaptive approach to AI governance, continuously refining its policies in response to technological advancements and practical experience gained from numerous AI pilot projects. The policy also supports the broader 2025-2026 AI Plan, indicating its role as a foundational document for the City's strategic AI roadmap.

International Alignment

While primarily a municipal policy, the Seattle Responsible AI Policy demonstrates an awareness of and aspiration for alignment with broader national and international principles governing artificial intelligence. The City's earlier Generative AI Policy, released in November 2023, explicitly aimed to align with the executive order on AI signed by President Joe Biden, indicating a commitment to national best practices and regulatory trends. This alignment suggests Seattle's recognition of the need for a harmonized approach to AI governance, even at the local level, to ensure consistency with federal directives and to leverage broader expertise and standards.

Beyond national alignment, the City of Seattle actively engages with various networks and coalitions focused on responsible AI. Seattle IT staff are involved with WaTech's AI Community of Practice, which works to provide guidance for AI development, procurement, and implementation by Washington state agencies. More broadly, the City participates in the GovAI Coalition, which is dedicated to using AI for social good, promoting ethical and responsible AI governance, fostering vendor accountability, and encouraging cross-agency collaboration. The City has also contributed to initiatives like the MetroLab Network's Generative AI for Local Governments Task Force, which explores how generative AI can improve city services while establishing policies for just, equitable, and accurate use. These engagements highlight Seattle's commitment to learning from and contributing to a wider dialogue on responsible AI, ensuring its local policies reflect current thinking and best practices in the evolving global landscape of AI regulation.

Implementation Timeline

Compliance Checklist

Sources and References