Page 38 - U4SSC Guiding principles for artificial intelligence in cities

IV. Assess results

This step involves either interim or final assessment of the results of implementing the guiding
principles for AI in a city. Cities are strongly recommended to conduct assessments comparing
actual outcomes retrospectively and objectively with respect to intended ones at the outset.

Similarly, the city can evaluate various enablers for their effectiveness during the implementation.
Gaps can be identified, in order to be addressed and corrected in due course. Lessons learnt can be
derived to understand positive and adverse consequences of the application of guiding principles
for AI in the city. Positive aspects of successful applications may be potentially cross utilized among
various AI solutions; for example, a successful policy in one AI solution may trigger the use of a
similar policy approach in another. Such examples can also be extended to other enablers. On
the other hand, identification of ineffective enablers would result in their potential relinquishment
in due course.

Guiding principles inevitably shape and transform AI solutions. Therefore, it is important to assess
their outcomes retrospectively. An ex-post impact assessment would be highly beneficial to
understand various social, economic and environmental changes that occurred in the city and
compare them with those intended prior to implementation.

The comparison of ex-ante and ex-post assessments will indicate deviations in terms of intended
and actual outcomes. Such deviations may aid in planning more accurately in due course or fine-
tuning guiding principles for AI.

28