In late 2014, two tech startups, Qentis and Cloem, presented their rather interesting business models, based on similar technologies, that could challenge the very basis of intellectual property laws as we know them today. Taking in part from its original mention and analysis at VentureBeat, this post discusses the models of each of the two startups, their basis and possible implications.
Picture a scenario where everything original anybody ever conceived and uploaded on the internet/released for public viewing resulted in a claim for copyright infringement. That rather staggering proposition is exactly what the Russia based Qentis Corporation has made. The company’s owner, Michael Marcovici, claims that his company has the ability to use massive computing power to facilitate the bulk algorithmic creation of words, music, images and almost anything else that is copyrightable on so large a scale, that by 2020, every possible photograph will have been pre-preemptively created and registered by the corporation, with its rights owned by the company’s clients, who apparently comprise private high net-worth individuals (HNWI), investment funds and corporations that act as pure investors.
Even with text, the company claims to be trying out combinations in groups of 400-word articles in, mind you, not one, but eight languages – English, German, Spanish, French, Russian, Polish, Portuguese, Italian claiming that there shall soon come to be a time when any literary work created will by default infringe upon the copyright holders to Qentis’ work. Qentis’ About page already states that the company is responsible for 97% of all feasible text that can be created in the above mentioned languages. In an attempt to establish their legitimacy, the company also claims here that Lady Gaga’s 2013 release Applause had already been computer-generated by them in 2009, where – save for the additional line here and there, the words remain identical.
The company seems to envision the establishment of a copyright monopoly of sorts, where the paths to all seemingly original creations will always lead to Qentis. Although this is fascinating news, it seems that the company would have to jump over some rather significant roadblocks to actually accomplish what it claims. The extent of computational force that it would require to generate all possible combinations of 400 words or less in a constantly evolving language like English, for instance, is magnanimous, thus, rendering it an extremely difficult task to undertake in five other languages as well.
Further, it is interesting to note that the US Copyright Office did announce that it “will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author”. However, circumventing this condition doesn’t seem to be a worrisome ordeal – placing an expert to make alterations to central elements within the algorithm at appropriate intervals might just about do the job. But the thought of doing the same for every algorithm produced, considering that millions are churned out every year looks quite the extreme opposite of feasible.
Additionally, Qentis’ apparent ability to generate text combinations cannot take away the rights of writers who have independently created original works of authorship by claiming copyright infringement, and considering that Qentis’ is highly secretive about what it creates and who possesses the rights to it, establishing that the creation was indeed independent shouldn’t necessarily be difficult at all. The converse would be true for Qentis as well – if Qentis had actually independently created the Lady Gaga song, it could legally compete in the market with Lady Gaga’s version of the song.
But really, how different from a computer churning out lottery numbers is a computer churning out different permutations and combinations of words? For all practical purposes, a massive portion of the work is performed by a machine – where is the creative value here that deems the work worthy of being granted a copyright? Occasional interventions made by an editor cannot possibly be considered adequate ‘creative input’ so as to enable circumvention of the restriction laid down by the Copyright Office. Originality will eventually become a nonexistent concept altogether if a version of every perceivable idea already exists.
Although Qentis is apparently an attempt at dry satire, listed along with other art projects at Marcovici’s site, which is again, funnily enough, called www.artmarcovici.com, it does well at driving the point home either way – near inconceivable IP law upending creations could very well be on their way.
Another other peculiarly fascinating tech startup is Cloem, which uses software language operations to manipulate a seed group of a client’s patent claims, for instance, by rearranging steps in a process, providing for the possibility of patenting of hundreds of new inventions. With Cloem, there is actual reason to believe that computer generated inventions may be patentable, considering that there have been a number of inventions in the past that have been wholly or partially conceptualized by a computer software.
Cloem’s website says that it combines human claim drafting expertise and computational linguistics to create cloems or ‘technical poems’, and makes use of over a 100 proprietary algorithms, specialized dictionaries and synonyms, hyponyms, hyperonyms, meronyms, holonyms, and antonyms to create these patent claims. In an increasingly competitive market, when a patent is granted, competitors may seek to acquire rights to close and neighbouring aspects of the patented invention, that may not hinder the implementation of the core invention, but it may prevent you from proceeding in a direction that you may have otherwise taken. Cloem may generate dozens of alternate machine generated claims that can act as prior art (provided they fulfil the requirements of a patent in that respective jurisdiction, of course) by anticipating similar future patent claims, and decimate its novelty by claiming rights to any whitespace surrounding the original patent, thus “cleaning up” the field and creating exclusive space for future patentable action for the original patent holder.
There are, however, certain concerns. Firstly, existing patent law recognize the inventor of an invention as a person or an individual only, and not a machine. Since the generation of cloems combines human effort as well as machine and software technology, it is questionable as to whether this combination falls within the ambit of inventor as defined today. Secondly, it is also debatable as to whether the Cloem’s software engineers and patent experts working on cloems pertaining to a particular patent can be considered joint inventors. Where an invention is the result of the efforts of multiple people, each person must have contributed to the patentable feature in that claim through the exercise of more than just ordinary skill. However, the original patent already embodies the core invention, whilst the engineers only partially contributing to writing a general claim code based on the original patent to establish prior art – does this really make the engineers co-inventors or does this comprise a situation where the original patent owner merely uses the services, ideas, and aid of others in the process of perfecting his invention without losing his right to a patent?
If the Courts do decide to accept the presence of cloems as prior art, this could have a rather undesirable effect on creativity and innovation. Patent holders will seek the cloems for their patents left right and centre, creating a rather unhealthy competitive practice that could discourage firms as well as individuals from creating altogether, with the standards for admissibility of a patent claim being pushed higher.
The coming about of Qentis and Cloem is a glaring indicator of how far technology has developed and how many more questions about our existing practices technology in the future can actually throw up – I can’t particularly deem their creation as undesirable, but it could really tip the playing field in favour of Qentis(if it ever came to be an actual, legit organisation), and entities that cloem their patents unless regulation gets to them first.