A man is rallying against patent trolls and the patent system itself by building a website that can mathematically generate new work.
The All Prior Art site, created by Alexander Reben, creates new work or prior art (evidence that an invention is already known) - making new patents harder to submit. Instead his site's randomly generated works are submitted for public use with a creative commons license.
He is looking to raise awareness of what he dubs the “antiquated patent system, and to preempt patent trolls” by scanning the entire published US manifesto of patents to algortihmically generate new work based on the strucutre, terminology and language of previous patents.
The project, spotted by Gizmodo is creating and publicly publishing "all possible new prior art, thereby making the published concepts non patent-able”.
The ambitious scheme often publishes nonsensical jargon in the hope of slinging together a functional design.
It's not a million miles from the infinite monkey theorem, that if you give enough monkeys enough keyboards, they’ll eventually give you Shakespeare.
Reben added: “While most inventions generated will be nonsensical, the cost to computationally create and publish millions of ideas is nearly zero – which allows for a higher probability of possible valid prior art.”
He admitted that the deep learning could be used to create future patents based upon previously submitted entries to the US Patent Office.
To date, 1,660,000 designs have been published by the page and can be fully downloaded.