Researchers at Stanford University claimed to have found that computer algorithms could accurately predict the sexuality of people from photographs of their faces. Their artificial intelligence programme was able to detect, with surprising accuracy, from more than 300,000 dating website images, who was gay and who straight. The computers were more successful at getting the sexuality of men right – 81% of the time from one picture, 91% after analysing five pictures. With women it was still impressive; 71% correct from analysing one picture and 81% after five.
There was a time when the powers that be had the ability to detect witches. For hundreds of years they used their talents to discover and then kill witches by the thousands and the tens of thousands. Then they discovered that witches didn't exist! "Oh dear. What a shame. Never mind," said the witch-finders. "But looking on the bright side; if all those witches were truly innocent then they went straight to heaven didn't they? So we actually did them a favour didn't we?"
More recently the authorities used scientific measurements to recognise criminals:
More recently the authorities used scientific measurements to recognise criminals:
A thief, for example, could be identified by his expressive face, manual dexterity, and small, wandering eyes. Habitual murderers meanwhile had cold, glassy stares, bloodshot eyes and big hawk-like noses, and rapists had ‘jug ears’ ... female criminals were far more ruthless than male; tended to be lustful and immodest; were shorter and more wrinkled; and had darker hair and smaller skulls than ‘normal’ women.
Given past results, I might just take this Stanford University theory with a grain of salt.
No comments:
Post a Comment