Responsible AI Strategy

By: City Detect Inc.

Updated 30 July 2024

As a firm that develops and deploys predictive AI solutions for municipalities, we recognize the importance of adopting a responsible AI strategy to ensure our technologies are ethical, transparent, and beneficial to society. This document outlines our dedication to responsible AI practices and provides a framework for implementing these practices.

Legality

1. City Detect’s imagery collection methodology is analogous to a local government official collecting an image of a property from the right of way using a photographic lens from a cellphone or digital camera. Katz v. United States, 389 U.S. 347 (1967), establishes that audio/imagery sensor captures in the public right-of-way are fair game and maintain no expectation of privacy. Gill v. Hearst Publishing Co., 40 Cal. 2d 224 (1953), establishes that imagery taken from the right-of-way is fair game and maintains no expectation of privacy. Aaron C. Boring & Christine Boring v. Google, Inc., W.D. Penn. No. 08-694, issued February 17, 2009, establishes that Google Street View imagery is also not an invasion of privacy. Additionally, City Detect trains its AI models on its own data and complies with all US Copyright and IP law.

Data Integrity and Privacy

1. Data Quality: Use high-quality, accurate, and relevant data to train our AI models.
2. Privacy: Protect the privacy of individuals by adhering to data protection regulations and implementing robust data anonymization techniques.
3. Security: Implement best-practice security measures to safeguard data from unauthorized access and breaches.
4. City Detect works alongside local government departments of Information and Technology to ensure its practices align with local standards.

Ethical Design and Development

1. Fairness: Ensure AI models are free from bias and provide equitable outcomes for all communities.
2. Transparency: Maintain clear and open communication about how our AI systems operate and make decisions.
3. Model Results: Ensure AI outputs that can be understood and interpreted by non-experts so that city leaders can spend their time making a difference, not sifting through data.
4. Model Explainability: Favor the simple solution – not the opaque one, ceteris paribus.

Continuous Monitoring and Improvement

1. Performance Monitoring: Continuously monitor AI systems to ensure they operate as intended and maintain high performance.
2. Feedback Mechanisms: Implement feedback loops to gather input from users and stakeholders to improve AI systems over time.