The Nightmare Machine

Deep in the bowels of the Massachusetts Institute of Technology, a machine is learning to scare us. The Nightmare Machine is actually a set of computer algorithms fed a diet of spooky images. It uses deep machine learning to find common scary features in images of haunted houses and graveyards, all the while searching for what we find terrifying. The next step is to show us. Here the machine applies the tricks it’s learnt, the goose-bump-inducing tricks of perspective, lighting and unease to Tower Bridge in London. But the Nightmare Machine isn’t finished – now it needs a steady stream of human brains. Visitors to this website can rank scary images, telling the machine which ones raise a whimper. Such interaction between human and machine may help to understand the psychology of fear and also suggest new avenues in artificial intelligence, which you may find even scarier.

Written by John Ankers

John Ankers

Writer

A former researcher in systems biology, John is now an online biology tutor and professional coach supporting the wellbeing of scientists and NHS professionals. He's writing, too, of course.

Website

• Image produced by Pinar Yanardag, Manuel Cebrian and Iyad Rahwan at MIT, from an original image on Flickr (CC BY 2.0)
• Massachusetts Institute of Technology, Cambridge, MA, USA
• Image copyright held by the photographer