Design trustworthy services
Government services can become contentious if they don’t verify that everyone receives fair and equitable treatment, and governments are expected to deliver value and benefits to the community. It’s critical that government services are explainable so they can be audited and appealed (as per administrative law and other requirements). It is important to verify that ethical considerations are part of the design, implementation and governance models for the service.
Government services also need to comply with relevant ethical guidelines in the design and use of all systems, especially for automated decision making (such as the use of artificial intelligence).
Questions to ask:
-
Do you know the relevant ethics frameworks and requirements for the jurisdictions?
-
Is your system designed to have privacy, dignity, legitimacy, and accountability by design?
-
Can an end user understand and challenge the output from the system?
-
How do you audit and monitor decisions, accuracy, and legal authority or compliance, in real time?
-
How do you know whether your service is having a fair, positive, or negative impact on people?
-
How have you verified that independent oversight and effective governance?
-
How could you earn social license and operate this service in a way that the public will consider trustworthy?
-
What are the jurisdictional requirements around transparency? For example, do you need to publish algorithms, or apply a special risk assessment?