[ttalle]

The problem is with the information inbalance. Before, a salesman, for example would look at what kind of clothing you were wearing and what segment items you're looking for.

Now let's do a little thought-experiment with the Uber ride and look at the information you and Uber have to make your price and choice on:

For Uber:

- What time is it? +

- Are you far away from home? +

- Have you been here before? -

- What have you paid before? +

- Where do you shop usually? +

- Are you rich? ++

- Are you ill? +

- Are you pregnant? +

- Are you drunk? ++

- Are you very drunk? +++ (bonus!)

- Are you sad? +

- Did you just break up? ++

- Do you have any friends? +/-

- Do you have any friends with cars? -

- Are you alone? +

- Are you going home? +

- Are you already home? -

- probably a lot more...

For you:

- Price?

- How screwed am I if I don't take the ride?

There are not many other options. Maybe a taxi's while they last and Lyft probably has the same algorithm.

[dredmorbius]

With machine learning, the model isn't even a set of explicitly selected factor whose parameters are then derived from data, but hidden factors (and interactions) within the neural net. The engineers who construct the models don't even knnoe how they achieve their ressults, only that they do.

Upside? Plausible deniability! An NN AI that happens to score based on age, gender, gender preference, religion, disability status, and ethnicity, but buried among other factors and not explicitly exposed? Perfectly fine! (At least until you can hire stats PhDs to prove statistically significant bias.)

Bad news? The engineers have no idea why (or when) the model gives bad results. It's answers without insight.

[hansvm]

Is the problem then that they have that information, and having that information is inherently bad for some reason regardless of the effects, or is it that with that information they're able to reliably capture more of the market surplus and cause more of your transactions to only be barely worth it for you?