A fascinating article in today’s Guardian about experiments Google have been running on their image recognition software. Basically they asked it to emphasize the things it most recognizes, and then they looped this to make the differences ever more extreme. So slowly things get quite surreal – check out the back of the knights’ saddle in the image opposite.
The software was also asked to make specific pictures out of screens of static, so we can see what it believes a banana looks like if left to its own devices.
Two things struck me. Firstly the images are pretty crazy. But they do give a sense of how a machines visual ‘minds eye’ looks. Which is very different to a humans (with the odd exception like Hieronymus Bosch or Van Gogh). But presuming the software will become more ‘intelligent’ over time, and able to generate images that are a true product of its ‘imagination’ we might start seeing wallpapers, branding and art that is unlike any aesthetic the human mind would naturally conjure up. So possibly we are standing on the brink of a whole new visual culture.
Secondly, when one hears about A.I. replacing call center workers etc. one is smugly reassured that its ok, the rest of us can be poets or designers. But perhaps we might be sharing office space with computerized creatives sooner than we might imagine?