Automatic generation of inventions for the public domain

In the field of patent enforcement, there is a well-known problem of patent trolls , which interfere with the work of normal companies, making absurd claims for "illegal" use of technologies such as " Internet advertising ", " network games ", " email scanning ", " bilateral and multilateral acceptance solutions "," system of assigning tasks to mobile workers "," system to do all sorts of things in different ways on a computerAnd many, many others. Trolls are very active: they not only use old patents of broad scope, as listed above, but also try to obtain new patents on technologies that are clearly in the public domain. Unfortunately, they often manage to trick the US Patent Office in such a way, which cannot cope with a thorough analysis of a huge number of applications. For example, the patent troll Rothschild Connected Devices Innovations in 2014 patented "remote control via the Internet . " The Foundation for Electronic Frontiers even holds a competition for the most stupid patent of the month , considering such cases, and actively fights with pseudo-patents .

New project All Prior Art- the original attempt to resist this madness. Author Alexander Reben has developed a program that generates descriptions of technologies - and publishes them online, which automatically makes technologies unpatentable.

Descriptions of technologies are drawn from the entire base of issued and unapproved patents, and are compiled. Phrases and sentences in the “patent language” are randomly intertwined. The author admits that most such inventions will be meaningless, but the cost of creating and publishing millions of ideas is close to zero. This means a high probability that at least some valuable thing will automatically fall into the public domain.

He expects that large organizations like Google can allocate servers and use more intelligent technologies like in-depth training to generate texts in order to “flood the public domain”.

Perhaps the combination of all possible words in a language will generally kill patent law as such in its current form. All possible ideas will already be in the database. Given technological progress, this is possible in the foreseeable future.

Currently 1,660,000 descriptions of inventions are available online ( current torrent ).

Descriptions are also duplicated on Archive.org , in volumes of 10,000 each.

For texts, a Creative Commons license is chosen to prevent commercial use of texts and derivative works.

PS Perhaps, only one thing threatens the existence of the All Prior Art project - all of a sudden someone has done this before and has already patented the idea ...