The 7 principles for responsible AI use in local government

Principles for the ethical use of AI

This year’s report looked at the biggest technology-related challenges facing municipal organizations. For 36 percent of respondents, determining the appropriate use of AI ranked high on their list.

To make sure AI serves the greater good, local governments may want to consider some guiding principles:

Beneficial to the public

Your AI systems should meet a clear community need. Just because AI can do something does not mean it should. Local governments may want to assess whether AI applications will be a genuine benefit to the community.

Accountability

People who design and implement AI systems need to be accountable for their outcomes. This includes conducting thorough examinations to understand the potential impact on individuals’ rights and well-being. Regular monitoring and audits can help make sure AI functions as intended and do not inadvertently reinforce biases.

Transparency

For the public to trust AI-driven services, they must understand what these systems do, how they work, and how decisions are made. Local governments need to clearly communicate about the technology’s role in service delivery, using plain and jargon-free language. It’s essential to set up protocols for citizens to challenge any AI-driven decisions that may seem inaccurate or unfair.

Fair and unbiased

AI systems should not create or reinforce biases. This means relying on high-quality, representative data and regularly reviewing algorithms to prevent bias from influencing future decisions. Municipal organizations need to ensure that human oversight is always part of the decision-making process.

Safety

Any new technology — including AI platforms — must be reliable and safe to use. Risks need to be continuously assessed and managed. Algorithms are designed to find patterns in data and can, at times, produce undesirable results that need human intervention. To identify errors and make necessary adjustments, local governments must have oversight mechanisms in place.

Privacy and security

AI systems must be designed with privacy in mind from the very beginning. Additionally, municipal organizations should ensure compliance with all applicable laws and regulations, as well as protect their systems from cyber security threats. This can be done through continuous monitoring and risk assessments to prevent unauthorised access and disclosure of sensitive data, ensuring system integrity and overall system availability.

Cyber threats aside, local governments must also consider the human element of security. Not all municipal employees need access to all data, and continuous security and privacy training — on topics like file sharing risks, password management, remote work security, and data handling and disposal — can make sure employees understand best practices and standards.

Values

Municipalities are people-focused organizations that aim to provide uninterrupted service to their citizens. The people, processes, and technologies involved in the design and implementation of AI systems should reflect the values of the community.

For instance, the values of diversity, inclusion, accessibility, and collaboration should be at the core of any AI system for a local government. A collaborative approach to AI is more effective in identifying and removing unfair biases as it encourages a more careful examination of input data, algorithm design, and its output. 

By adhering to these principles and focusing on privacy and cyber security preparedness, local governments can responsibly implement AI to improve public services while safeguarding the data and well-being of their citizens.

Privacy and cyber security preparedness

As more and more local governments implement AI, the security implications grow more serious. AI systems can create new vulnerabilities, making it more likely for cyber criminals to exploit sensitive data or gain unauthorized access to municipal infrastructure.

To address these risks, local governments need to implement AI systems with privacy and security at their core. This means conducting formal privacy impact assessments and implementing safeguards to protect sensitive data from unauthorized access or breaches. Regular audits and continuous monitoring can help identify suspicious activities and vulnerabilities before they are exploited.

But as digital engagement with citizens increases, how prepared are local governments from a security standpoint? The answer: there’s still some work to be done.

As per our municipal report, only 14 percent consider themselves to be very prepared. Fifty-seven percent are somewhat prepared, while 19 percent are not very prepared. Four percent are not prepared at all.

Interestingly, AI can also play a role in improving your cyber security posture. Advanced tools can detect unusual patterns or activities that may indicate a threat, enabling local governments to respond quickly and mitigate risks. These systems can also streamline risk management processes by predicting and addressing vulnerabilities before they become issues

Is your local government ready to responsibly implement AI?

Here’s the thing about AI — it’s likely already being used by your municipal employees. Whether there is a formal policy in place or not.

And there’s good reason for it. AI systems have become an essential tool for local governments. However, it’s a tool that needs to be wielded responsibly. Here are some best practices your local government may want to consider ahead of an AI implementation:

• Ensure strategic alignment and a value-based approach to identifying and prioritizing AI initiatives
• Assess the impact of AI-based systems on citizens and business owners
• Conduct privacy and risks assessment before implementation of AI systems, and monitor those risks and any emerging behaviours as they get used
• Be clear about business ownership and accountabilities
• Be transparent about how the AI-based systems work and their outcomes, and test and monitor them regularly
• Ensure they are safe; establish contingency plans and procedures; exercise human oversight allowing for human intervention when appropriate
• Ensure data quality, free of bias
• Provide adequate employee training
• Consult with your organization’s legal services early in the project to ensure compliance with legal requirements