>A Delta spokesperson told Fortune the airline “has zero tolerance for discrimination. Our fares are publicly filed and based solely on trip-related factors like advance purchase and cabin class, and we maintain strict safeguards to ensure compliance with federal law.”
Convenient that all the discrimination will happen in an AI black box so no human has to take responsibility for the outcomes, or even acknowledge that discrimination may be occurring.
I wonder if destinations, or cabin classes, or meal preference or whatever other "trip related" data they collect is strongly correlated with any protected classes.
But humans already take bullets for improperly using the AI, why should it be different here? Here's an AI hallucination cases database - I know I know it's only hallucinations, but discrimination is also illegal (for the time being at least) https://www.damiencharlotin.com/hallucinations/
That list seems to be focused on parties showing up to court with AI-hallucinated information.
The majority involve false quotes and citations. These are verifiable things. A citation reference can be followed and it will either refer to a real case or it doesn't, and then the citation either matches what was said in that case or it doesn't.
Proving that the ticket price AI discriminates in an illegal fashion would be much more difficult in court, especially when there is no longer any basis for a "normal" ticket price.
It results in the same kind of situation as when someone is not hired because of their race, but the hiring manager would never say that. They'll construct some other reasoning. How do you prove what was in their head when they made the decision?
Convenient that all the discrimination will happen in an AI black box so no human has to take responsibility for the outcomes, or even acknowledge that discrimination may be occurring.