Ethics are relational: A Critical Systems Thinking approach to implementing ethics in medical AI

Magali Goirand* 1, Elizabeth Austin1, Robyn Clay-Williams1

1AIHI, Macquarie University, Sydney, Australia

 

WHY: 

Implementing ethics in medical AI is complex, and needs to be transparent, contextualised and inclusive.

Methods: 2 phase study including interviews and focus groups

Results: 

37 people, 

23-83yo, average 52yo, 

19 women, 18 men

Conclusion: questions to investigate:

How affected are human relationships by the introduction of the AI?

What is the capacity of the agents of the system to handle the information from the AI?

How transparent is the process of custodianship of the data, algorithm?

How does the purpose of the app align with the purpose of the stakeholders?

What are the values embedded in the AI and how explicit are they?

Who can stop the program if it does not fulfill its purpose?