
In season 3 of Westworld (yes, I know no one watches it anymore), there is a powerful AI called Rehoboam.* This AI can predict world events and manipulate the lives of individuals to create peace and harmony in society.
Rehoboam not only predicts the events of people’s lives, but it makes assessments of their value and what kind of life they will live.
A main character in the season, Caleb, is unable to find stable employment to improve his lot in life because Rehoboam intervenes in his job search to limit his options since it has determined that is what role he should play.
I’m just going to leave that to sink in for a moment.
The AI pegs individuals into certain streams and then limits them to those streams by manipulating their environment despite their desire for something more.
I was reminded of Rehoboam’s role in the Westworld narrative during a conversation about AI in HR, which started with the obvious benefits of AI in being able to parse skills and competencies and assess candidates as a result. The potential to reduce bias through focus on skills is one of the possibilities offered through skills focused analysis.
But then the conversation turned to some of the risks, including the risk that systems might peg people a certain way (the examples provided were HR systems and college admissions systems).
Of course, Rehoboam is a fictional character in a fictional world, but these kinds of stories are often our way of working through some of the potential ethical dilemmas we could face. If you don’t believe me, think of all those movies and novels about clones right about the time we mapped the human genome.
Throwing the baby out with the bathwater** with regards to predictive AI isn’t going to avoid the risks. But it is interesting to think that the same potential that AI offers to eliminate bias through skills focus might also be matched by its ability to amplify bias.
*a problematic name as it is also the 4th king of Israel and successor to Solomon, known for his wisdom
**an odd phrase, apparently German in origin, dating back to the 16th century
Leave a comment