Three-year-old Ursula is hit by a self driving Toyota Silph and killed. The car's object-recognition software correctly identified her as a human child, but prioritised not hitting an endangered bird.Â
The system was trained using data from many sources, including pupillary analysis that identifies what people actually care more about.Â
Other than that white band, it’s exactly like any other woodpecker, but because of that white fucking band it has four Wildlife Preservations Acts. Four, which is four more than the number of acts dedicated to regulating weighted risk matrices in autonomous vehicles.
They’re smart enough to read your email and measure your pupils and listen to your phone calls, they have access to all of the data on who we are and what we love. They’re smart enough to understand how much a mother loves her baby girl. They’re smart enough to understand the emotional impact of killing a woodpecker. They’re smart enough to know what they did and they’re smart enough to keep doing it, right?
(..) the neural network training for cultural understanding of identity is collected via social media, keystroke analysis, and pupillary response to images. They’re watching to see what’s important to you. You are responsible.
| Title | Publication Type | Year | Creator |
|---|---|---|---|
| STET | Narrative, Short story | Sarah Gailey |
