AI governance Archives - City Innovation Network

Tag Archive for: AI governance

Dublin unveils strategy for responsible AI adoption

23 June 2026/in Jonathan Andrews/by Folkert Leffring

As Ireland’s capital city prepares to launch its AI strategy, Jonathan Andrews spoke to Jamie Cudden, Executive Manager for Corporate Services & Transformation with Dublin City Council, a founder member of the City Innovation Network

Dublin City Council has established what it describes as Ireland’s first Generative AI Lab dedicated to local government, creating a structured environment where staff can test artificial intelligence tools, develop practical use cases and build organisational capability before wider deployment across council services.

Through a combination of formal oversight groups, mandatory staff training and controlled experimentation, Dublin is seeking to balance innovation with public trust, transparency and compliance as AI becomes increasingly embedded in local government operations.

An AI strategy for Dublin City Council is scheduled for publication in the third quarter of 2026 and will set out the organisation’s vision and principles for deploying and scaling AI.

Jamie Cudden, Executive Manager, Corporate Services & Transformation, Dublin City Council, says the objective is to realise the benefits of AI while ensuring strong governance and accountability.

“We are focusing on human oversight, transparency, privacy, fairness, accountability and security from the outset, rather than treating governance as something added later,” he says.

Building governance before scale

Dublin City Council has introduced a two-tier governance structure to oversee AI adoption across the organisation.

An AI Governance Group provides strategic oversight and decision-making around policy, acceptable use, organisational readiness and priority use cases. Alongside it, an AI Technical Group focuses on implementation matters including infrastructure, tool assessment, permissions, security, cyber risk, technical evaluation and support for pilots and use cases.

The groups bring together representatives from IT, Digital, Legal, Data, HR, Corporate Services and Transformation.

Jamie Cudden, Executive Manager, Corporate Services & Transformation, Dublin City Council

“Coordination happens across business, innovation, information systems, data protection and cybersecurity functions so that AI proposals are considered from multiple perspectives before they move forward,” says Cudden.

The council’s approach is informed by the Irish public service framework for responsible AI and by the requirements and direction of the EU AI Act.

Prioritising use cases

The council evaluates AI proposals according to business value and organisational need. Potential projects are assessed against questions such as whether they solve a genuine service or operational problem, can be implemented safely and align with wider organisational priorities.

Particular focus is being placed on use cases that improve internal productivity, strengthen decision-making, enhance customer service and support better use of data.

“We ask: does this solve a real service or operational problem, can it be implemented safely, and is it aligned with wider organisational priorities?” adds Cudden.

He says proposals are considered through a governance lens rather than a novelty factor.

“This helps ensure that AI activity is linked to broader business and service delivery objectives, rather than being run as standalone experimentation,” he says.

A central part of Dublin’s strategy is the GenAI Lab, launched in February 2025 in partnership with the ADAPT Research Ireland Centre and Trinity Business School. The initiative is described as Ireland’s first Generative AI Lab dedicated to local government.

The lab functions as a sandbox environment where council teams can work with researchers to identify opportunities, test ideas, develop prototypes and build organisational capability.

Rather than allowing unrestricted experimentation, the lab operates within a structured framework where activity is logged, reviewed and supported by governance and technical oversight.

“The emphasis is on learning by doing, but within a framework that protects the organisation, staff and the public,” explains Cudden.

The lab is being used to identify practical opportunities across council services while helping staff develop familiarity with emerging AI capabilities. Researchers and council teams work together to explore where generative AI can add value while ensuring projects remain grounded in governance, ethics, transparency and public trust.

Guardrails and responsible deployment

The council has developed staff guidance on generative AI and introduced governance structures and controlled environments to evaluate potential use cases.

Areas of focus include data protection, information governance, human review of outputs, transparency, security controls and awareness of bias and error in AI-generated content.

The council is also placing emphasis on the use of approved tools with appropriate safeguards rather than consumer-grade platforms.

“Our guardrails are built around the principle of responsible use before scale,” he says.

Workforce development forms a significant part of Dublin’s AI programme.

The council has published AI guidelines for staff and introduced an introductory AI training course which is mandatory for employees. The 80-minute training programme provides foundational knowledge on AI opportunities, risks and responsible use. More than 500 staff have completed the training to date.

Dublin is now preparing a broader programme that will target more than 3,000 staff, alongside more focused training and development packages for specific roles and functions.

“The intention is not just to give people access to tools, but to help staff understand where AI can be useful, where the risks lie, and how to use it responsibly in a public service setting,” Cudden explains.

The GenAI Lab also supports workforce development through workshops, engagement sessions and hands-on experimentation, allowing employees to gain practical experience while exploring potential use cases.

Collaboration and peer learning

Cudden says engagement with other cities, research institutions and professional networks helps the council compare governance approaches, share lessons from pilot projects and benchmark its progress.

“Collaboration plays a very important role in our approach. AI adoption in local government is still evolving, so peer exchange is extremely valuable in helping cities learn from each other, compare governance models, share lessons from pilots and avoid repeating mistakes.”

Dublin has engaged with cities including Sunderland and Madrid and participates in city-to-city initiatives such as the City Innovation Network.

The council also works through research partnerships and smart city networks to test ideas and contribute to wider discussions around responsible AI in urban government.

“Dublin has been active in engaging with other cities and networks on digital innovation, and we see this as essential to developing a practical and credible approach to AI,” he says. “Our involvement in initiatives such as city-to-city exchanges, research partnerships and wider smart city networks helps us test our thinking, benchmark progress and contribute to the wider conversation on responsible urban AI.”

Moving from pilots to enterprise deployment

The council is developing a pathway from pilot projects to operational deployment, including assessment criteria, governance checkpoints, support models and resourcing arrangements.

“At present, the focus has been on putting in place the foundations for responsible experimentation, governance and capability-building,” says Cudden. “The challenge we are facing now is scaling the pilots into enterprise solutions. The proof of concepts are delivering quick and impactful results across multiple use cases. We are scoping out how we support enterprise infrastructure that allows us to scale applications within a safe and secure, sovereign environment to utilise the latest AI models and tools that are GDPR compliant.”

The AI strategy due at the end of 2026 is expected to formalise the council’s vision for AI deployment and scaling.

“We need to be fully open and transparent in this regard so we can build trust with our staff, councillors and the public that we serve.”

Dublin’s AI governance model

Governance structure

AI Governance Group responsible for policy, acceptable use, organisational readiness and prioritisation of use cases
AI Technical Group responsible for infrastructure, tool assessment, security, permissions, cyber risk and pilot support
Representation from IT, Digital, Legal, Data, HR, Corporate Services and Transformation
Governance informed by the Irish public service framework for responsible AI and the EU AI Act.

GenAI Lab

Launched February 2025
Joint initiative between Dublin City Council, ADAPT Research Ireland Centre and Trinity Business School
Described as Ireland’s first Generative AI Lab dedicated to local government
Sandbox environment for research, testing, prototyping and capability building
Supports identification of AI opportunities across council services
Activity logged, reviewed and supported through governance and technical oversight.

AI guardrails

Staff guidance on generative AI introduced in May 2025
Human oversight of AI outputs
Data protection and information governance controls
Transparency requirements around AI use
Security and cyber risk assessment
Consideration of bias and AI-generated errors
Focus on approved enterprise tools rather than consumer-grade platforms.

Workforce development

Mandatory introductory AI training programme
80-minute foundation course
More than 500 staff trained to date
Future programme planned for more than 3,000 employees
Additional role-specific training and development packages under development
Workshops and hands-on learning delivered through the GenAI Lab.

Next steps

AI strategy scheduled for publication in Q3 2026
Development of enterprise AI infrastructure capable of supporting AI applications at scale
Focus on secure and sovereign deployment environments
GDPR-compliant deployment model
Formal pathway being developed from pilot projects to operational deployment.

Main image: Alex Grichenko | Dreamstime.com

Seattle sets new benchmarks with responsible AI plan

7 October 2025/in Jonathan Andrews/by admin@cityinnovation

Seattle has launched a new Responsible AI Plan and Community Innovation Hackathon Series designed to strengthen how the city governs artificial intelligence while embedding fairness, privacy, and community engagement at the centre of its approach.

The plan expands policy beyond generative tools to cover all AI solutions, establishes clear rules for prohibited and high-risk uses, and commits to hiring a dedicated AI lead who will oversee a new cross-departmental AI Governance Group.

Rob Lloyd, Seattle’s Chief Technology Officer, told Cities Today that the new approach is designed to move beyond pilots and ensure projects are aligned with the city’s most pressing priorities. He explained that proposals must now demonstrate sponsorship and value and will be closely monitored to ensure they deliver results.

Rob Lloyd, CTO, City of Seattle

“There will be initiatives where things do not produce, but the goal is to catch them early and to fail forward for future work, protecting security, privacy, and resources,” he said.

Balancing innovation with the workforce

Lloyd stressed that the city is not rushing into AI adoption without preparing employees.

“The City of Seattle began with training and upskilling before rolling out general access,” he said. “Our intent in that Fundamentals, Approaches, and Solutions structure is to support the people and change management parts of what AI is causing our employees and teams to see.”

The city is prioritising AI for tasks that often go unfinished or add strain to staff, with the goal of reducing workloads and supporting better decision-making. Employees and labour partners are helping identify areas where AI can provide the most value, from accelerating insights to addressing work that otherwise cannot be completed—all without displacing jobs.

Measuring success and engaging communities

Seattle also wants to ensure historically marginalised groups shape how AI is used. “Through participation, outreach, and engagement. It invests in positive impact in and for these communities,” he said. “And by watching the impact and adjusting based on what we see. Seattle is a national leader in our community engagement work, including through the leadership of our Department of Neighborhoods and our Innovation and Performance Office.”

The new AI Plan includes a Community Innovation Hackathon Series in partnership with AI House, bringing together students, community members, technologists, and city staff. Its first hackathon theme, “Youth Connector,” is focused on using AI tools to improve awareness and access to mental health and enrichment programmes for young people.

The city will track outcomes against key measures such as performance, reliability, adoption, scalability, return on investment, and alignment with business goals. Lloyd said these metrics, combined with stronger governance and training, will help ensure AI is deployed responsibly and with lasting impact.

Seattle was one of the first cities in the US to release a generative AI policy in April of 2023, which was then iterated and fully implemented that October. The original policy established governing principles for the city’s approach to AI including: innovation, accountability, reliability, fairness, privacy, explainability, and security.

Main image: Gregor Doerr | Dreamstime.com

Denver expands CIO role to include AI

3 October 2025/in Jonathan Andrews/by admin@cityinnovation

Denver has become one of the first US cities to formally embed artificial intelligence into its top technology leadership, with Suma Nallapati’s role expanded from Chief Information Officer to Chief AI and Information Officer.

The new title reflects the city’s ambition to be seen as the nation’s most forward-thinking AI city and to ensure that technology is deployed in ways that are equitable, secure, and centred on residents.

“Implementing AI across city government inherently presents a few risks–bias and equity, transparency and accountability, and security and data protection,” Nallapati (pictured) told Cities Today. “My strategic approach aims to mitigate these risks by diving in with eyes wide open.”

Nallapati pointed to the city’s AI Request for Proposals as a major milestone. The initiative is designed to create a pool of trusted vendors capable of delivering secure, scalable, and innovative solutions tailored to Denver’s needs.

Following a strong response, the city is now reviewing submissions against criteria such as technical strength, compliance with security standards, scalability, cost, and innovation potential. Each vendor must also complete a detailed risk assessment, which is scrutinised by Denver’s security and privacy teams.

The city’s most visible deployment so far is Sunny, a multilingual chatbot supporting residents via web, text, and WhatsApp. Between January 2024 and September 2025, Sunny engaged with more than 102,000 residents, answered nearly 100,000 questions, achieved a 90 percent satisfaction score, and absorbed 30 percent of daily 311 interactions.

“I believe AI’s success is ultimately defined by how much better it makes life for our teams and the people we serve,” Nallapati said. “With tools like Sunny, we’re tracking the number of residents engaged, our customer satisfaction score, the percentage of daily interactions that are being absorbed by Sunny that would have historically been routed through our 311 call centre, and more. And, since Sunny can recognise 72 different languages, we’re also ensuring the service is accessible and equitable for every Denverite.”

Her wider strategy focuses on transparency and collaboration with an approach to AI that is grounded in active engagement and collaboration.

“We are not just planning transparency; we are practicing it,” she said. “Last year, we successfully launched workshops to boost AI literacy among our city agency, who also contribute their expertise to policy review through our Information Governance Committee.”

Nallapati said Denver is drawing on national best practice by working with universities and foundations, including participation in the Bloomberg City Data Alliance, which is helping shape ethical guidelines and resident data initiatives. At the same time, the city is expanding its existing feedback systems so that residents can directly raise questions about AI, making public input a formal part of how new tools are deployed.

Collaboration with vendors will continue but without ceding control.

“We absolutely need the expertise of partners, but Denver will maintain control and accountability,” Nallapati said.

The city will retain 100 percent ownership of its data and will audit the algorithms, and the city’s in-house team will ensure vendor systems meet standards for ethics, transparency, explainability, data protection, and security.

Nallapati added that Denver aims to set the national pace by embedding AI across every layer of government, from data governance and policy to frontline services.

“Our mission is to seamlessly weave AI into the resident experience,” she said. “Our ultimate goal is to become the blueprint for AI-enabled cities where technology actively enhances equity and access.”

Other US cities and states that have already added, or plan to add, significant AI roles to their workforce include Louisville, New York State, Montana and New Jersey.

Image: City and County of Denver