Put your intelligence to the ultimate test and see how you fare compared with other people.
In around half an hour, the artificial intelligence developed by a team at Imperial College London – nicknamed Cognitron - will put you through a series of a dozen customised tests and, after you have supplied a few details, tell you how well you did.
The online study uses web-based AI developed by neuroscientists Romy Lorenz, Rob Leech, Pete Hellyer and Adam Hampshire at Imperial’s Computational, Cognitive and Clinical Neuroimaging Laboratory, ‘C3NL’. The AI will harvest information from thousands of people and dozens of tests, enabling it to explore hundreds of different measures of cognitive ability.
The subject of intelligence remains contentious, not least because there is still no agreement on precisely what the word means. Tthe AI will tinker with the tests to find out if intelligence can be divided into different types of cognitive ability like verbal reasoning and focused attention, or if such cognitive skills are all interdependent. Ultimately, Cognitron aims to understand if AI is the key to understanding human intelligence.
The more people who take part in the AI experiment, the richer the insights into human intelligence that should emerge. Cognitron should be able work out the major components of human intelligence by harnessing the ability to learn in an iterative manner. According to Adam Hampshire, if the test reaches more than 20,000 participants, ‘the AI can learn and map out the structure of human intelligence in greater detail than any previous study.’
Also involved in the test with Hampshire, Lorenz and Leech are Pete Hellyer, Wellcome Trust fellow at C3NL, William Trender and doctoral student Ricardo Pio Monti, who studies with mathematician Giovanni Montanna from King’s College London.
Find out more about the project in the October issue of WIRED magazine, on sale now. The feature story,The AI Science Machineis also available online.
This article was originally published by WIRED UK
