“We may never get there [patenting] advanced machines if we do not patent machine learning processes. ”
The idea of patented inventions brings to mind machines that are fully realized – flying devices and engines with gears and pistons that work in coherent symphony. When it comes to artificial intelligence (AI), there are no things, no gears, no pistons and in many cases no machines. AI inventors sound much more like philosophers who theorize about machines, rather than mechanics who describe a machine. They use phrases like “predictive model” and “complexity module” that evoke little or no imagery or coherence with practical life at all. The ways of the AI inventor are in contrast to the principles of patent writing, where inventions are described in relation to what does what, why, how and how often.
But our patent laws were written for machines and widgets, not this predictive modeling. It’s hard to patent AI because AI is not a machine – it’s a brain for a machine. And the brain has to learn in order to be trained before it can be used on a machine. The machine philosophers who design AI? They teach the brain. Clearly, the U.S. Supreme Court does not want us to teach that brain. Machine brains do not exist in their eyes, only machines.
American innovation is being left behind
The rest of the world sees things differently. Unhindered by similar legal constraints, China is advancing in the AI development sphere, developing exam guidelines specifically for big data processing and mathematical formulas, the tools to build advanced machine learning, and deep learning capabilities. Europe requires only a technical purpose for the patentability of computer-implemented inventions, rather than a technical improvement.
In the United States, big data processing and mathematical formulas are practically non-patentable topics in themselves, being considered as abstract ideas that lack the technical improvements needed to overcome Alice 2A / 2B analysis. Technical improvements require improvements in the way the machine works, which is the standard of both the United States Patent and Trademark Office (USPTO) and the US Court of Appeals for the Federal Circuit.
The problem is that almost all machine learning and deep learning inventions will be considered non-patentable under our current laws because machine learning cannot survive the current technical improvement analysis. In radiology, advanced imaging, which can detect early-stage cancer from ultrasound, depends on huge amounts of data sets of images fed to a machine for processing – data sets that take into account age, ethnicity, body type and a myriad of confusing factors. which can affect the image. Eventually, the machine will be smart enough to take into account all the confusing factors and be able to identify critical issues for the healthcare provider. For autonomous vehicles, the AI inventor must teach the brain everything – literally all road conditions for all possible navigation scenarios, to have an autonomous vehicle brain that does not propel you into a bullet pit. Advanced artificial intelligence, capable of predicting and making decisions, is built using big data processing (which relies on cloud computing) and mathematical formulas (coding). We need cloud computing and coding to build advanced neural networks. The courts consider all these processes to only implement a generic computer to act in a conventional way.
The problem is that our legislators assume that denying the patenting process to the machine learning phase is not equivalent to cutting off innovation in machine performance. There seems to be a notion that we can bypass the phase of big data processing and mathematical formulas, but once we get to the good machines – the really smart ones that can find cancer – yes, we can patent those.
This assumption is flawed because we may never get to the smart machines for the following reasons:
1) The Commercialization Problem: Effective machine learning modules need to be implemented in commercialized products to be viable. In the case of autonomous vehicles, AI machine learning modules must first be implemented in existing non-autonomous vehicles. We can not expect to have a programmer deliver scenarios to the autonomous vehicle brain and expect the brain to be good. The brain has to face the unquantifiable scenarios from the real world of road navigation in order to safely direct the vehicle to make that left turn at the next light.
It’s no secret that companies compete with each other and they do not want to reveal their inventions through commercial use, just to have it pledged by their competitors, especially inventions that will require many years of research and development investment such as building a smart self-propelled machine. Denial of eligibility discourages these types of investments and further downstream investments.
2) The common collaboration problem: Radiologists do not code! Even if they do, they are probably bad and incredibly expensive. But radiologists understand imaging. When radiologists collaborate with programmers and coders, together they can build machine learning modules for digital labeling of vast amounts of imaging data needed to develop advanced imaging AI in healthcare. We need joint collaborations between members of different technology sectors to build and implement AI across different sectors. These joint partners want to prevent the establishment of ownership of their output. Our Act on Eligibility for Topics may discourage joint collaborations by inserting a blind spot in these agreements.
We may never get to advanced machines if we do not patent machine learning processes. We are witnessing the emergence of the fourth revolution’s greatest enabling technology in artificial intelligence, which will affect all the major technological spheres from cleantech to advanced telecommunications. But even now, the Supreme Court is refusing to rectify the shortcomings of U.S. law American shaft. The Supreme Court has made it clear that they do not intend to change course with regard to their views on American innovation policy. Unless Congress takes the lead, we in the United States should prepare to become the country where imagination and machine philosophy remain just that – mere ideas.