Page 17 - U4SSC Guiding principles for artificial intelligence in cities

• Obtaining consent from individuals (the information provided during the consent may differ in
terms of providing details for its purpose, potential impacts, and the way data are processed);
• Providing means to control use of data;

• Redressing mechanisms (e.g., providing means to modify and delete data); and

• Incorporation of privacy during AI systems design rather than as an afterthought, i.e. “Privacy
by Design”.

3.2.3 Fair and Inclusive

AI systems make decisions or automate certain tasks on behalf of human beings. AI algorithms
and the underlying data which AI systems use (e.g., for training purposes) may potentially include
biases. These biases may stem either from AI systems design or in some cases from intrinsic or
historical biases existing in organizations operating these AI systems. Figure 1 provides examples
of various biases that contribute to harm within the data and AI application processes.

7