How can Machine Learning help us make moral decisions?
How can Machine Learning help us make moral decisions?
Presentation to “Responsible AI” Meetup group
The new wave of machine learning hype has sparked even more interest in the possibility of using statistics and data to make important decisions— especially decisions about other people. In this talk, the speaker will first retrace how this recent phenomenon is not even new — for instance, psychologists have known for nearly 70 years that statistical diagnosis is more accurate than human diagnosis. Given that psychologists still predominantly use non-statistical methods to make their diagnoses, the speaker will ask why that is the case, and what lessons we can draw for moral ML decision-making generally. The speaker will conclude with some pointers to the different challenges, obstacles, and limitations faced by the current push to use ML to make moral decisions about people.