According to whistleblowers Google’s AI isn’t “neutral” but reflects the biases of the designers. SHOCKEDSHOCKEDSHOCKED!!!!!!
Jesus Christ. You don’t need a whistleblower to “reveal” this. Every real programmer or record-keeper or accountant knows that a program reflects the preferences of the designer.
This is especially true of models and simulations, no matter how complex. Any predictor, from a spreadsheet to an AI simulation, NECESSARILY follows a bias and serves a purpose. There are an infinite number of available variables in the world, and an infinite number of ways to incorporate and weight each variable. It’s intrinsically impossible to bring in EVERYTHING in an unweighted way.
Even if you take everything available on the web, you’ve already selected an EXTREMELY NARROW piece of the world. What makes it to the web is a 2-dimensional shadow of reality, planted by a tiny subset of people. Most people don’t write any quantity of text or images. The people who postest the mostest are INFLUENCERS, who are selling an idea or a product BY DEFINITION.
And what makes it into Google’s searches is vastly narrower and biased-er. It already reflects Google’s ideology, as anyone who has ever run a search knows.
Polistra's Mill 