Page 18 - U4SSC Guiding principles for artificial intelligence in cities

Figure 1: Examples of Biases in AI Applications (adapted from Source: NIST)

Additionally, these AI systems may not be inclusive in terms of the broad target potential group for
which they are designed. It is important to ensure that no one is left behind in the target potential
group, and that geographical, economic, social and digital equality is preserved by including all
relevant people with different gender, nationality, ability, age, ethnic origin, and so on (including
all marginalised and vulnerable groups).

Hence, this principle allows cities to purchase, deploy and use unbiased, non-discriminatory, and
inclusive AI systems.

Implementation Considerations: Cities can adopt various mechanisms to boost the fairness and
inclusiveness of their AI systems. These mechanisms include:

• Developing governance procedures around the procurement of third-party technologies
incorporating AI technology;

8