Artificial intelligence might seem like it’s still a faraway science fiction concept, but the technology is already here. It’s not yet at the level of mimicking human behavior and thoughts, but it’s regularly used in research and development. Acquiring data is one area where it is invaluable. Big data has become a very valuable resource that is well worth acquiring. Researchers and marketers are currently investing in technology that makes use of big data. However, there are a couple of issues that arise out of using data created by others, especially when AI is involved. Here are some ways this technology affects IP rights.
Consumer data and AI
AI systems can make some big leaps when they are able to use big data. They are often supplied to businesses that seek to utilize their consumer data and create a higher demand for their products and services. The underlying technology platform is often trained to find set patterns among consumers that are relevant to the business. It’s a lot more complex than it sounds.
The results that the AI produces can be considered a grey area when it comes to ownership. Who owns the IP rights of the customer-specific analysis? There are many examples where this comes up as a problem. There is a bit of a controversy surrounding AI cybersecurity systems. They detect network intrusions by comparing current consumer traffic with some baseline traffic from the network. Who actually owns the database or copyright of these sets of data?
The issue becomes more complicated when consumers start switching suppliers. They take with themselves the analytical data that was generated by the AI. How do you prevent them from reusing this data? The copyright law isn’t very clear when it comes to this topic, but you can be sure that the original contract will come under very close scrutiny.
Competition and privacy
Despite the many positive aspects of big data and AI, there are many skeptics that would take this technology with a grain of salt. There are many outspoken voices about the negative effects of this little industrial revolution. The data-rich markets of today will gladly take in more data when given the opportunity.
Big data is considered a boogeyman of privacy. Some people believe that concentrating all of this information within a couple of powerful hands might not be the best decision. The data might get manipulated and used for nefarious purposes and manipulation. This can be a cause of great market failure and it might give a bit too much power to the companies that control the information. There are even fears that big data can be used to circumvent democracy. Something along the lines of gerrymandering is a lot easier when you know your constituents well enough and their behavior becomes predictable.
To protect the privacy of consumers, several laws have been enacted by different governments. Data dominant companies are required to pay a data tax . This forces them to share their information and knowledge with smaller competing firms. The EU takes its privacy very seriously. They have released their General Data Protection Regulation to give the EU some clear goals when it comes to privacy and big data.
Originality and AI
For something to be copyrighted, it needs to be considered original. However, what makes something original, anyway? It has to be the author’s original creation and expression of their creative freedom. If the author puts their personal stamp on the creation, you can pretty much call it original. It’s not too strict of a definition, either. The threshold for originality is pretty low according to copyright law.
Even though copyright has a pretty broad scope, it still requires a human touch. Intellectual human intervention is required for something to truly be considered an original piece of data. This is why raw data like forecasts and sports scores still don’t fall under the umbrella of copyrightable material. Unfortunately, there’s no clear consensus on what falls under copyrightable data, which is why there debates regarding this very topic.
The eligibility of protection for a certain set of data points needs to be reviewed on a case by case basis. It should be done in light of the rules of the country where the topic is brought up and within the broad scope of case law.
Presentation
Many experts agree that some part of big data can and should be copyrighted. Despite it coming from AI and not from a human mind, these kinds of data points are very important and need to be protected. Where they come from is considered less important than how they are presented.
A big part of big data protection is the presentation of the data points. You might not be able to copyright AI-generated data on its own, but you can certainly copyright information that is presented in an original way. Data that is connected and analyzed isn’t original enough. This is why those that deal with big data on a daily basis will shape the data in a specific way. By expressing this knowledge in different forms, you can call it an original idea that is worth being copyrighted. When it comes to the legal side of things, this tactic works flawlessly. It overcomes the legal ambiguity of copyrighting the data by adding an original human touch to it.
Infringement on existing data
AI and training systems can occasionally infringe on third-party copyrights. If part of the AI training process makes use of third party data, the end product could be considered an IP infringement. After all, it’s using someone else’s copyrighted resources.
The artificial intelligence program is the one using the data, but it can’t be prosecuted. The AI system is not a real person, which means it can’t incur liability. Therefore, the person controlling and using the AI may be liable for copyright infringement. On the other hand, AI is often given a bit of leeway when it comes to which data sets they use. As long as the source is relevant and useful to the search, the AI will try to use it to create the best possible end result.
Many liability challenges can arise out of this. How do you figure out whether or not the AI will infringe on third-party data? More importantly, how do you differentiate stealing a certain dataset and coming to the same information through research? The truth is, there’s no reliable way to see if the AI is intentionally set to find third party information. This is why there are still many debates about the legality of AI research.
Inventiveness and evaluation
The waters get muddy when you use AI to invent new ideas and products. The use of AI systems in research and development is not a new concept, but the legal aspects are still heavily disputed. It adds a layer of complexity to an already difficult to assess the issue.
Imagine an AI that is designed to create new material for use in construction. It is fed as much data as possible about all the different materials that are available to the business. Using this data and some input parameters, the AI could create an optimal new material for the business. However, the way the AI does this poses a problem.
Company A has built the AI system and designed it to work a certain way. The AI is then able to find a new material with the right inputs. Company B puts the AI to use in a particular engineering context and adds inputs. The AI comes up with four different viable materials. Company B uses the first option to build a car roof and discards the other material ideas. Company C analyzes the four materials and finds that the fourth one could be useful for their own products. This company then copyrights and incorporates the material in their new products or services. The end product that company C uses is arguably the result of inventiveness and practical knowledge, thus they are the rightful owners. Although, an argument can be made that both companies A and B have made contributions towards creating the material.
This is a common dispute when it comes to AI and copyright. Pharmaceutical companies run into this problem every once in a while. Identifying new drugs is done via high throughput screening, which is a patented method.
Working on fixation
For a piece of work or information to be copyrighted, it needs to be in some concrete form. You can’t just decide to protect abstract data, especially when it’s sourced by artificial intelligence. It has to be transformed into something tangible.
There are several ways to go about this. Data can be translated into several different forms. The simplest form would be writing a handwritten note. You end up with a piece of paper that has worth, but it’s not a very practical solution. There’s no practical reason to turn enormous amounts of data into a bunch of handwritten notes. Plus, handwriting is pretty unreliable.
Photographic documentation is another option. Having visual evidence of the data is a great way to fixate it. However, in most cases, this kind of data is handled in the form of digital files. This way, they are easily accessed, shared, and analyzed by professionals.
Rights and exclusivity
Copyright holders are granted many exclusive economic rights that help them control a copyrighted work’s use. Copyright law helps protect the owner’s data from third parties that would use it without any authorization. The most important rights granted by copyright law include the ability to reproduce, communicate, and distribute the data. When balanced with copyright exception, this allows for the most optimal kind of protection for the copyright protection holder’s data.
There are certain problems that can arise in the context of big data projects. In data environments, it’s necessary that you obtain authorization from the owner of each individual data set. This would require finding and seeking authorization from hundreds, if not thousands, of different works. In many cases, it’s difficult to assess the degree of authorization that the copyright holder has given you and your business. Before the data can be used, a thorough analysis needs to be performed.
Protecting the acquired data might require legal assistance in some cases. Third parties will very likely try to abuse the grey areas of big data protection. To prevent misleading & deceptive conduct disputes from affecting your projects, you’re going to need to find adequate legal protection. A big part of abstract data protection is handled in the courtroom.
Trade secrets
Copyright and database rights help businesses take control of their works. They provide measures for enabling control over the diffusion of works and ideas. This is especially important for data that fulfills the originality criteria. Protecting trade secrets and ways of acquiring data is important for a business. It allows the copyright owner to keep their commercially valuable assets safe from the competition that would take advantage of it. This makes it worthwhile to invest in acquiring the data in the first place, doing so without worry that someone else will use this data without investing the same kind of effort.
Trade secret protections have their own advantages. When it comes to big data projects, trade secret protections act as a safeguard for individual pieces of data that might not be entirely clear on the originality front. What also helps is that it doesn’t differentiate between different kinds of data and their applicability. As long as the information has not been disclosed, trade secret protections are unlimited in time.
Conclusion
IP laws have not yet caught up to the newest technological improvements that are used in product research and development. This has led to IP disputes where it isn’t clear who is in the wrong, or if any infringement has occurred. One thing is certain, big data will have a huge impact on IP laws in the coming years.
