From Plato to Yoda, training responsible AI designers for on-the-field action
Friday 23 October, 2020at 12:15 - 13:00
Would Plato have been so virtuous if he had a mortage to pay and a project to finish by Monday? The caricature of a "responsible AI designer" is often someone good at convoluted philosophical discussion, at always finding something to complain about and at standing for his or her greater principles – no matter the cost.
However, evidence shows that, unlike pure philosophical ethical principles and theoretical thinking, day-to-day ethical decisions arising from living human minds are relatively bendable, particularly when external factors are at play, such as hierarchical pressure or threat over one's status.
As a matter of training the responsible AI designers of tomorrow, we need to consider all the skills that are needed for responsible AI design – beyond the one of the ivory-tower ethicist.
Here we will investigate the less-expected yet very-critical Jedi skills that are needed for real-world on-the-field responsible AI design, such as psychological resilience, self-defense against manipulation, and convincing the audience that some droids are "not the droids they are looking for" – even the gold-shiny one. What are these skills? How to develop them? How to assess they have been acquired?