#Google is well-known for its enthusiasm for artificial intelligence, such as the #AI-powered Google Assistant loaded into the company’s new Pixel smartphone. As part of this quest to make ever-smarter computer programs, Google today unveiled its new AI experiment, called Quick, Draw! Taking the form of a rapid-fire online game, Quick, Draw! gives the player six objects to draw, using only black lines, with a 20 second time limit for each drawing. A neural net AI constantly attempts to figure out what the drawing is meant to be. Google hopes that the AI will eventually learn from the doodles until it can guess 100% of all drawings.

Neural nets explained

Neural nets are designed to mimic the structure of the human brain. Each net is composed of multiple, highly interconnected processing units, which react dynamically to inputs. Instead of using a single, powerful central processor, like those used by IBM’s super-smart Watson AI, a neural net relies on many more very simple, highly-connected processors. By assigning greater priority to certain connections in response to specific stimuli, the AI can ‘learn’ how best to respond to a range of inputs.

The slow march of artificial intelligence

Quick, Draw! isn’t the first time a tech company has tried to train an AI to recognise visual inputs. In the spring of last year, #Microsoft launched an age-guessing bot, which could attempt to work out the age of an individual based on a photograph. This was followed up by an automatic caption bot, which would work out the content of an image and write an appropriate caption to go with it. This feature was incorporated into Microsoft’s infamous millennial twitter bot, Tay, which was quickly tricked by trolls into describing German dictator Adolf Hitler as "swag".

Just like Microsoft’s experiments, Quick, Draw! hasn't quite reached the level of artificial intelligence seen in film and TV just yet. Twitter users have been venting their frustration with the inability of the vaunted neural net AI to recognise some very basic, obvious items, such as smiley faces and swans: