Artificial Intelligence: Why AI Inventions Will Disrupt Patent Law

Artificial Intelligence (AI) has taken innovation and creativity to the next level. It has hastened advancements in technology using data, algorithms, and computing power. AI tools are being used by companies and individuals as “innovators”. In the 20th century, AI was a remote thought in human civilization and development. It was just assigned to science-fiction and had remote chances of being put into reality. Today, AI has advanced to the level of offering various new impacts on the lives of men. AI has helped most nations to improve their healthcare, production, and commerce, money transfer, education, rescue, energy generation among other processes. AI has improved social welfare of man.

Movie Industry Prepared the Human Race for AI

Most people know about man and the machine from watching films based on artificial intelligence. They just know that it is fiction, but the technologies are becoming real with the advancement of AI. Machines have been showcased in most sci-fi films as disruptive to the normal life of man. Arnold Schwarzenegger films Terminator Series 1-4: The Rise of the Machine and the ensuing series show machines in full force. Other films such as Tron, Blade Runner, The Matrix, Star Wars, Metropolis among others have embraced artificial intelligence and shown a concept of man vs the machine. The films show machines as doing extraordinary work with or without input from man.

Contention in Patenting AI Inventions

One issue is that AI’s technological advancements have gone to the next gear such that they will upset the legal frameworks of patent laws of countries such as the United States. In the history of patent law, the system was created to offer incentives to innovation by offering exclusive rights to inventors for a short period. This was to be done for a limited period in exchange for their inventions. These laws were designed without the knowledge of AI in the picture. Several dilemmas stemming from current laws across the developed nations are set to crash with AI inventions.

The main contention stems from patent applications and the fact that many nations have laws that cannot be amended to give copyrights to non-humans.

The Three Main Parts of Patent Laws That Are Set To Be Disrupted

A white paper presented by Baker McKenzie at the World Economic Forum Center for the Fourth Industrial Revolution offered critical insights to the three major issues surrounding patents of AI inventions.

1.     The patentability and Inventorship of AI inventions

The American Patent Act has not specifically demanded that the creator of an individual must be human. But, the American constitution defines the personality of an individual which locks out AI created inventions. One of the most common concerns about AI inventions is about protecting them. Patent law should adequately cover the inventions that have been done by the AI technologies.

2.     Infringement liabilities caused by AI

In most nations such as the United States, they fail to recognize the liabilities involved when someone infringes on AI inventions. The patent system only recognizes the human innovator with an intellectual output from the mind. This exception sets AI inventions to the public for infringement.  When AI inventions go wrong, who will be held accountable? Is it the computer that engaged in machine learning or it is the person that triggered the creative spark?

3.     Eligibility of the patent subject matter of AI technologies

Given the acceptance of AI technologies by most people, the legal standards for patentable subject matter is of concern. Currently, there is much tension surrounding AI inventions. This means that given the ethical and social benefits of AI, there should be ways of changing the patent laws to embrace the new AI inventors.

Artificial Intelligence Patents

In artificial intelligence machines are creators. AI is a subject matter of several patents founded on inventions in machine learning, deep machining, natural language processing, data analytics and natural language search. AI is automating current resources and data to generate new inventions for businesses.

Companies Investing in AI Patents

Top notch companies such as Uber, Apple Inc. Google, and Tesla are employing data scientists that can use machine learning to make AI inventions. These companies have invested in AI laboratories spending billions of dollars annually to make the next inventions that will improve the welfare of man. In 2016, these large corporations spent a total amount of $39 billion on AI investments. The revenue from AI is set to grow from $8 billion in 2016 to more than $47 billion in 2020. As AI inventions get adopted by businesses and organizations, the journey to patenting AI inventions keeps gaining pace. In 2012, the number of AI patents granted was 708. In 2016, there number rose to 2,888. In the United States, almost 1,628 patents were issued during the four years.


Three Examples of AI Use in Production of Original Books, Songs, Paintings, Poems and Music

In the past, man was the creator of patents. Man’s discoveries, innovations, and researches were done and patented. Today, machines are creating original things that are also being patented. Today, computers are editing and writing poems, and photographs.

·        Google

Some of the Google AI patents have been filed under the company DeepMind. DeepMind is an AI program that will author and publish news articles. This AI outfit has made software that it uses to create new music through listening to past songs. Between November and December 2016, DeepMind filed more than 10 software patents. Other enterprise AI companies that Google is seeking to acquire include Banter, AI Matter, Halli Labs,, Moodstocks among others. Google intends to use these companies to produce AI patents that will push its innovation to the next level. Currently, Google has about 14,500 AI patents

·        Microsoft

Microsoft, the tech giant has the most AI patents in the world. Microsoft has applied for about 16,800 patents.

·        Ebay

Ebay has AI patent publications assigned in machine learning algorithms, data processing methods, SEO (search engine optimization) and neural networks. These AI patents have been applied in the running of Ebays search engine optimization, inventory management, management of online resources and studying the behavior of Ebay users.

·        IBM

IBM has embraced AI patents on communication technologies, neural networks, machine learning algorithms, AI market analytics and computer-implemented method that analyzes the data for images. The AI patents assisted IBM to track resources that are online, image identity and processing, online shopping, monitoring vehicles in real time and analysis of traits based on posts made on social media by users.

·        Amazon

Amazon has AI patents in cloud-hosted application resources, computing protocols, data processing techniques and machine learning algorithms. These AI patented systems assist Amazon in client surveillance. It has also helped them to establish a recommendation system for offline retail businesses.

·        Book Authoring in Japan

In 2016, a computer program in Japan authored a book that almost won the national literary prize.

·        Production of Paintings in Netherlands

A Netherlands program known as The Next Rembrandt was made by researchers and groups of museums. The program analyzed all the pieces of art done by Rembrandt van Rijn. The recognition algorithm had studied data from the 346 known paintings in a period of 18 months.  After that, the program is expected to draw the next pieces of art just like the way Rembrandt himself would do.

Critical Questions on AI Inventions

In the case of the Google’s Deep Mind, software; who then takes the credit for creating the song? Is it the computer that has uses machine learning and deep machining or the person who invented systems that created DeepMind? What happens to the original composers and producers of the songs and books that DeepMind used?

Patent Law Dilemma in the United States

Patent laws in the United States were strong before the year 2000. Most innovators and investors rushed to register their creations under the USPTO. In the last decade, several legal battles have been litigated owing to the amendment of the old IP laws and administrative processes. Some businesses and companies took advantage of patents leading to fines for businesses that had violated IP laws. Some laws limited innovation in science especially biology and software. Patents were outlawed and even applying for new patents was an issue. The patent specific court, the Federal Circuit has had its rulings appealed to the United States Supreme Court. The court has got hold of patent damages and violations are difficult to prove.

In American IP landscape, it is difficult for patent holders to prove that their patents have been used illegally. Besides court proceedings, administrative processes have been created by the Patent Office of the United States. The introduction of the Inter Partes Review (IPR) has changed the number of patents being filed. In the past six years. About 7000 patent related cases were filed in the Federal Circuit. More than half of them are related to patents based on software while the rest is based on electrical, biological and chemical patents. IPRs are used to reduce costs and risks associated with the patent litigations. Patent infringement defendants have used several instruments to protect themselves against the companies that are hiding behind the “Fair Use” policy.

 Is AI an Inventor?

Owing to the increased number of cases on patents based on software that have been generated by AI processes, you will notice that AI is a force that should be watched closely.  Today, software is being created by AI technologies and they are making products and services that have not been seen before on the planet. There are more than 22,000 AI patent applications in the United States alone. The other positive signal is that most people are accepting AI inventions meaning that a debate to accept AI as an inventor is foreseen in the near future.

Definition of an Inventor According to the United States Patent Act and Laws

The definition of an inventor in the American patent law makes all the difference. An inventor is defines a person who discovered an invention. One of the legal gaps is that the American Patent Act fails to define and individual. In USPTO manuals, an inventor is defined as a person who conceives the invention and puts it into practice. Therefore, AI is not a human entity because it does not have the mental nature of conception and putting the thoughts into practice and second it is not an individual. The US Patent Laws thus limit the protection of intellectual property that has been invented by AI because the technologies are not proven to come from an individual.

Let us assume that AI has the power of conceiving an idea, making an invention and reducing it into practice using machine learning. Why can’t the AI process be given a patent? In the US Patent Act, an individual must be the inventor. The same law provides that an inventor must be an individual of the Homo Sapien family. In the last proceeding in the Federal Circuit, natural persons can be “inventors” which throws AI technologies off the current discussions.

How Other Nations Treat AI Patents

American Patent Laws

American laws on patents have a predetermination that an individual is the one that has the authority to receive a patent. The dilemma arises if a human who preconceived the AI process that created the invention can get the patent. As computers become inventors, the laws as set by the American Constitution are limiting the patent owners. The problem lies when other nations have amended their laws to recognize computers as patent owners. The lack of recognition by the United States law that AI technologies can invent means that most inventions are limited.

Canada’s Patent Laws

In Canada, the patent law is different. The inventor has neither been defined nor restricted by the law. In Canada’s Supreme Court, a ruling was made in Apotex Inc. V. Wellcome Foundation Ltd., 2002 SCC 77.  In the case, generic drug manufacturers argues that the patents were not valid if some of the parties that led to the inventions were excluded. In the ruling, Canada’s Supreme Court defined an inventor as someone who preconceived an idea that is new and useful in the process, machine, tools that were used to make a manufacturing concept or process. Just like the American patent law, the inventor has not been defined. The Canadian patent law is not yet computer friendly but the fact that it recognizes processes, art, machines and tools used in the invention procedure, it is close to accepting AI technologies as an inventor. Currently, even Canada does not have the legal authority to assign AI technologies the human factor that allows it to be called a patent. In Canada, the inventors must be human too which puts it at the same pace with the United States.

Europe’s View on AI Patents

In the CJEU (Court of Justice of the European Union) a landmark ruling on C-5/08 Infopaq International A/S v Danske Dagbaldes Forening revealed that copyright is preserved only for original works. Originality must be a fruit of the author’s own intellectual creation meaning that machines have been thrown out as inventors. This ruling implied that for a patent to be given; the personality of the author is paramount otherwise, there is no copyright.

Australian IP laws on AI Patents

In Australia, the Acohs Pty Ltd v Ucorp Pty Ltd ruling stated that a computer invention cannot be protected by copyright law because it was not produced by a human entity.

Countries that Give Authorship to the Programmer

The United Kingdom, Ireland, India, Hong Kong and New Zealand have patent laws that recognize the author of the program which was used to generate the invention. The patent laws in those five nations offer an exception by recognizing human authorship needs. The laws also recognize the work of the programmer that generated the creative spark leading to the invention.

Scaling the Future for AI Patents

The Creativity Machine by Stephen Thaler generated an invention that was given an American patent.  The same case applied to the Invention Machine by John Koza. Regardless of the current state of the Patent Act in the United States, they may have to comprehend and embrace the increasing AI inventions and offer patents. The law may need to be reformed to include AI inventors to speed up innovations. The world must equally notice the downside of AI inventions which may lead to atrophy of human intelligence. The film, Cyborg illustrates the need to regulate AI patents.  It should be a balanced field for the human inventors and the AI inventors.

In the race to patent AI technologies and associated invention, China has taken the lead. China amended their Patent Laws to allow publishing of AI, Deep Learning, Machine Learning patents. The number of patents in China has risen at a fast rate in comparison to the United States. A report by Zhen Fund and Sequoia Capital China reported that America has 35,508 AI patents while China has 34,345. Currently, China and the United States hold half of all the AI patents in the world. For instance, America owns 32% of the patents in machine learning and 265 of the natural language patents. China has 23% of the machine learning and 14% of the natural language. China and America have different focuses on the AI fields that they specialize in. For example, China has 55% of the patents on machine vision patents such as facial recognition. The United States giant corporations such as IBM, Microsoft and Google have focused on speech recognition, synthesis, and machine learning patents.

There is a stiff competition from China in terms of the AI patents. Chinese patent laws seem to be a bit more flexible than the American Patent Act. The United States designed patent laws before AI, Machine Learning; Deep Learning were invented. The University of Surrey has published new research in the Boston College Law Review about granting patents to computer inventions. The research recognized that many companies are investing in creative computing but the current Patent Act does not recognize computers as inventors. A change in the law is needed to enable companies profit from the potential of creative AI technologies. Ryan Abbot, a Law and Health Professor at the University of Surrey’s School of Law concluded that naming computers as inventors is an incentive that would foster the creation of intellectual property through promoting the development of creative computers. Professor Abbot adds that while patent prosecutors continue holding to the fact that computers are not inventors, AI has produced some great ideas that have been used for decades, for example, “The Creativity Machine.” In the future, computers will be responsible for almost all the inventions. The creative machines will be smart professionals in their fields and therefore they will have unlimited knowledge that will help them make innovations.