Tempe adopts ethical AI policy to get ahead of risk

The City of Tempe in Arizona has adopted a policy on the ethical use of artificial intelligence (AI), following City Council approval. It aims to balance the opportunities of AI for innovation, efficiency and improved resident experience with risks such as bias.

The city-wide policy states: “The City of Tempe is committed to designing, developing and deploying AI technologies in a responsible and ethical manner.

“We recognise that AI has the potential to significantly impact society and drive innovation, and we believe that it is our duty to ensure that its development, adoption and use align with the principles of fairness, transparency, accountability and respect for human rights.”

Tempe is believed to be the first city in the state to formulate such a policy and have it enacted by the council. It reflects a growing push from some cities to establish guardrails for AI.

Unintended consequences

Nikki Ripley, a spokesperson for the City of Tempe, told Cities Today: “Tempe has not yet deployed novel AI technologies such as predictive data analysis or automated review of employment applications to gauge qualifications. The city’s current use of technology, including AI, does not extend beyond everyday or off-the-shelf varieties.

“Our hope is that by implementing purposeful, advanced reviews of specific AI technologies before they are used by the city in the future, we will be able to guard against potential unintended consequences including possible discrimination or bias.”

The policy applies to technologies such as predictive analytics, machine learning, deep learning, generative AI, and automated decision-making.

It stipulates that departments should clearly state the problem the AI technology would solve including the purpose and the scope of application.

Transformational potential

The city will “take steps to prevent discrimination, protect privacy, and mitigate the risks of unfair or unjust outcomes resulting from AI algorithms.” It has also pledged to clearly disclose where individuals are interacting with AI or AI-generated content.

Tempe’s Technology and Innovation Steering Committee will establish standards for accountability and oversight of AI and develop reporting mechanisms for AI-related activities. The committee will also define consequences for non-compliance with the policy, including potential disciplinary actions or termination of contracts, and establish remedial measures to address failures of AI systems.

“AI technologies have the potential to offer exciting new capabilities at the City of Tempe. Predictive and generative technologies are features of our phones, computers and entertainment,” Ripley said. “Before adopting AI technologies with such transformational potential, it is important for us to thoughtfully evaluate what the impacts may be on our community.

“One of the guiding principles that will drive this approach is the involvement of people in the ongoing use of AI technologies as well, including through measures like proactive quality control checks to ensure that AI use over time continues to be accurate and unbiased.”

The policy was developed using resources available through the What Works Cities Certification programme, run by Bloomberg Philanthropies and Results for America to help cities make better use of data.