This article is my short essay for Lund University and the course AI & Law.
It’s main objective is to highlight the mutually benefited interaction between intellectual property and artificial intelligence with empasis in copyright which lies at the core of any AI system.
Prologue
The legal landscape for artificial intelligence and intellectual property law is constantly evolving. That refers to all intellectual property creations generated to any notable extent by AI systems, including patents, trademarks and copyright works. The legal landscape of innovation and technology is continually progressing to tackle the multifaceted legal issues brought by the integration of AI systems into our daily existence.
We must draw attention from the beginning that intellectual property law is central to AI, particularly since copyright safeguards the software that animates and instructs the AI system.
Copyright and IP lies at the heart of AI since it reflects the base upon which the AI system will teach itself and finally make the decision for anything that is asked to do so. The data that will feed the AI system with the information needed to consume for its functions can be chosen by the programmers who can direct the system to neglect or emphasize in certain sources excluding some others. These are human interventions to the AI systems that envisage the programmers choices. The AI system can think brilliantly and autonomously but the framework and the directions are given by the code and the programmers.
Moreover, the decision making process, the criteria upon decisions will be made, what will be considered as acceptable and what not in the decisions tree, all these that define the potential results in an AI system operation, are programmers decisions and directions given to the AI.
So the copyright is inherent to all AI systems when we think about their fundamental elements -software, hardware, and the extensive datasets they rely on. The foundational element of everything lies in the code, mapping out the directives for the system’s self-training journey—deciding which elements to incorporate or exclude and establishing the priorities that govern its evolution in decision-making. And the code is copyright protected. AI software, just like all types of software, is generally protected by copyright law, including the source code, object code, and the overall structure, sequence, and organization of the software.
Of course there are certainly many more complexities to address in the context of copyright law and AI, as well as within other dimensions of IP law, like patents and trademarks. The legal system is currently working to navigate and resolve these challenges.
In the initial phases of the evolving relationship between Law and AI, it is expected that the various jurisdictions will exhibit distinctions and differentiations in how AI systems intersect with intellectual property.
We can anticipate however that similar to previous technological advancements, the three primary government entities -legislatures, agencies, and courts- will establish common basic doctrines and principles to regulate AI and IP, across the various jurisdictions.
The global integration and international expansion of both AI and intellectual property will, over time, foster shared understandings and practices concerning the legal aspects of AI and IP law. Given this, we can assess the current international landscape surrounding today’s AI systems and intellectual property law with emphasis on copyright, which lies at the core of Artificial Intelligence.
I. AI & Copyright
In the context of copyright, there is an ongoing debate about whether a digitally composed original work f.e. a poem, an image, a book, a drawing generated by an AI system qualifies as an original work.
Most jurisdictions today including US, EU, UK, and Canada, consider that original works are exclusive to human creation and not to machine creation. Consequently, works not authored by humans are generally excluded from eligibility for intellectual property copyright protection.
In alignment with this principle, even if an individual provides directions or inputs to an AI system, leading to the creation of a copyright work, they are not recognized as authors since the work is machine-generated.
The legal framework however will change to adopt the new reality.
There is a potential argument that a jointly authored work, combining human direction and AI system generation (e.g., a digital work), could be classified as a joint work under copyright. This would acknowledge the contributions of both the person guiding the AI and the owner of the software code driving the creative process.
Let’s examine the case of a digital picture being created from an AI software under the direction of a person. Who will be considered as the author from the perspective of intellectual property, the human user of the software or the intelligent software that executes its instructions and creates the painting? The correct answer is both.
The work can be considered composite, with the two independent and autonomous contributions of each contributing to the final creation. For example, let’s say a person has the idea to use speech-to-image conversion software, creating an original work depicting astronaut Neil Armstrong offering a bouquet of flowers to Leonardo da Vinci’s Joconda on the moon. The machine borrows from the big data information in its database, incorporating depictions of Neil Armstrong in his astronaut suit, Joconda, flowers, and a lunar landscape. The final creation partially satisfies the user, who intervenes by providing further corrective instructions to the software.
“Make Neil Armstrong kneel to Joconda as he offers the flowers. Change the lunar landscape from black to brown. Add in the background of the image the Earth, and make the flowers appear different from earthly ones, as they could be if they bloomed on the moon.”
The artificial intelligence software of machine learning executes the user’s instructions, creating a new image pattern, which again may undergo corrections, interventions, and creative instructions from the user, to be executed again by the software. Eventually, the produced work can be considered original and artistically sound, creating an image that can be commercially exploited as a digital creation sold as digital posters or in print.
Throughout the entire process, it can be argued that both the software user and the software creator decisively participated in generating ideas and following the user’s instructions. Both contributions could be considered equal in contributing to the final result, making both the user and the software creator co-creators and co-owners of the final work from the perspective of intellectual property.
In conclusion, it is crucial to emphasize the contribution of artificial intelligence in efficiently overseeing unauthorized reproductions of original works exploited by external parties. Leveraging the ability to rapidly scan the internet, AI systems can pinpoint cases of unauthorized use and distribution of original works, equipping creators with enhanced resources to protect their rights.
II. AI & Patents
AI related inventions refer to innovations that involve collaboration between humans and AI systems and others that are generated by AI systems alone.
Most jurisdictions have a clear answer about inventions being generated by AI systems alone as non eligible for patent law protection. The United States patent office, the European patent office and the UK intellectual property office, do not allow AI systems to be listed as inventors. They accept only natural persons to be considered as such.
Of course the answer is different if an AI system is being used as a tool under the directions of the inventor to contribute to novelty or to other parts that a product or process may enjoy according to patent law. In these cases AI technologies are just a part of the whole invention which is being conceived, directed and executed by a person who is legitimately considered as the inventor instead of the intelligent system.
However, according to Tim W. Dornis the emergence of artificial intelligence suggests that we are approaching the conclusion of patent law as we know it today. AI is able to “invent” not only new materials and machines but also manufacturing processes, pharmaceutical drugs, and household products. Soon, he states, our life will be replete with artificial artifacts and humans will no longer stand at the center of the creative universe—we are no longer the masters of innovation.
The central question regarding patent law and artificial intelligence revolves around the contributions of the creator and the machine to the final product.
If the innovative elements of the patented product or application rely mainly on human creativity, with the use of artificial intelligence simply being a tool under and subjected to the human guidance and inventiveness spirit, patent law will attribute the status of the creator to the person.
On the other hand, as we can see in the following graph on the right side, the more the artificial intelligence machine constitutes the main and central factor of producing novelty in a patent, the less likely it is that the final product will be protected by intellectual property law.
It is a matter of evidence and a matter for national authorities to examine each case and be ready to recognize a patent, and for the courts to ultimately decide on the criteria by which they determine whether the human or the intelligent machine played the key role in creating a product that seeks recognition as a patent.
III. AI & Trademarks
The interaction of Trademark law and Artificial Intelligence is clearly the one with the most mutual benefits among the other forms of Intellectual property
When it comes to trademarks, the law focuses on distinctive elements -such as branding, signs, or symbols- that can differentiate the products or services of one enterprise in relation to others and the competition.
Trademark law applies and confers its rights and protection to the eligible recipients, be they natural persons or legal entities, without regard to the method of branding. So from trademark’s law standpoint it is indifferent whether the distinctive trademark was created by AI artist systems or by human agents.
Of course trademark law is benefited by other applications of AI systems.
AI tools can be employed to enhance the efficiency of trademark searches. Automated search algorithms can quickly analyze vast databases to identify potential conflicts with existing trademarks.
AI technologies, such as machine learning algorithms, can be utilized for brand protection purposes. AI systems can monitor online platforms, social media, and e-commerce sites to identify instances of trademark infringement or counterfeiting and can continuously monitor various channels for potential trademark infringements.
On the other hand, the rise of AI applications could raise challenges in relation to trademark law. AI-generated content could pose challenges for trademark owners when the intelligent systems will be utilized to create logos or brand elements from well known and established designs that resemble existing trademarks but are not identical, leading to potential exploitation of foreign reputation for the legitimate trademark owners.
It’s important to note that the field of AI and trademarks is continually evolving, just like the other areas of AI and IP law but generally speaking trademarks will mostly benefit by the rise of AI technologies.
Epilogue
Artificial intelligence will continue to rapidly evolve in all areas of human activity and creation in the coming years. Intelligent machines will be a key ally in the creation of intellectual property and patents, serving as an unexpectedly powerful ally for every creator. Gradually, they will be able to autonomously and independently create almost anything. However, at the end of the day, it is the code upon which all intelligent systems rely and human choices that will always provide the framework and accepted options through it for the operation of all intelligent systems.
This doesn’t mean that AI systems will act positively, lawfully and always for good ends. Their outstanding intelligence will execute upon their programmers provided framework of choices. They will offer the results that were asked and programmed to provide.
It’s the responsibility of societies and law enforcement systems to safeguard that detrimental operations and unlawful results from humans directing and controlling the AI technology, will be tracked promptly and dealt appropriately.