So, can you copyright an algorithm? The answer depends on what you consider an algorithm to be. Is it a Law of Nature, a Trade secret, or a Sui generis idea? Here’s a look at some of the legal issues related to algorithm copyrighting. Once you decide on the legal status of your algorithm, you can start searching for a patent. If you’re looking for a patent for an algorithm, there are a few steps you need to take.


  1. is it a Trade secret?
    • Below are some important considerations for algorithm owners.
  2. Patenting algorithms
  3. Sui generis

is it a Trade secret?

Many companies classify algorithms as trade secrets. While trade secrets do not create exclusive rights to know-how, they do protect proprietary information from illegal acquisition and commercial use. This is an important mechanism for ensuring that the secrecy of algorithms is maintained and that third parties cannot use them. In many cases, algorithms are valuable business secrets and should be protected as such. However, this protection does not come easy.

Below are some important considerations for algorithm owners.

  • The disclosure must be limited. If you only disclose the algorithm to a select group of people, it could be considered an infringement of your trade secret. Regardless of whether a trade secret has been protected by copyright laws, a third party may be required to follow certain rules in order to avoid violating your copyright. While disclosure is not always illegal, it may defeat the purpose of enforcing the intellectual property protection.
  • Taking legal action is a serious matter. In some cases, a person must pay a large fine and spend five years in prison if the trade secret is stolen. However, if the secret is less than $100, it may only be considered a misdemeanor. Therefore, copyright is an essential step in protecting an algorithm. But what about if someone accidentally shares it? Is it okay?
  • The computer program must provide a substantial economic advantage to its owner. It must also be unique and have some sort of competitive advantage over other users. The plaintiff’s computer system obtained revenue by licensing the program to others. Thus, the plaintiff is entitled to compensation if they are infringed upon. So, when the computer code is used, it will no longer be a trade secret. And if someone copies the program, they are likely to violate the copyright law.

The GTSA also provides for several remedies. Injunctive relief, damages, and attorneys’ fees may be awarded. Damages, if awarded, may include the actual loss of the plaintiff or any unjust enrichment benefit the misappropriator received from the infringement. However, damages can never be more than twice the amount of the loss or the unjust enrichment benefit that the misappropriator gained from misappropriating the information.

As a result of the lack of transparency in the AI system, the trade secret law has an enormous impact on AI research. It hinders access to algorithms and explanations behind automated decisions, reducing transparency throughout the system. And the duration of a trade secret is unlimited. This could potentially mean that AI systems will remain opaque for decades. Ultimately, it is in the best interest of society and business to move AI systems to an appropriate IP framework.

Patenting algorithms

A rubik’s cube algorithm

In a recent case, the Federal Circuit held that the addition of binding arbitration language to a legal document is an unpatentable mental process. However, the court remanded the case to lower courts to consider whether the method of performing such an action automatically by a computer is obvious. Then, the examiner will assess whether the claim is reasonable, the references cited, and the arguments presented by the inventor. If the algorithm passes all these tests, the patent is granted.

The argument against the patentability of an algorithm is simple: all algorithms are directly describable in mathematics. That’s the same argument for chemical processes. Math and atoms are two separate things, and so computer simulations of Tylenol atoms would not infringe a Tylenol patent, since it’s math and not a cure for aches and pains. However, mathematical formulas can be patented if they are used for business purposes.

Computer algorithms are often referred to as software in the patentability discourse. Nevertheless, computer algorithms are generally considered to be patentable subject-matter. Moreover, software and computer programmes are often used interchangeably as a term for patented subject-matter. Nonetheless, there are a few important differences between the two. While algorithm-based computer programs are more similar to each other, algorithms may be less obvious than computer programmes.

AI algorithms may not be patentable. The only way to ensure their emergence is to develop the appropriate legal framework and establish the patentability of their components. This is possible through copyright protection for software and patent licences for algorithms that classify data. The EU has made this a top priority. But how can AI-based innovation be protected? The US Patent and Trademark Office has issued guidelines for machine learning. This guidance should help all of us make the right decision when it comes to patenting algorithms.

The first three steps of a hypothetical algorithm do not fall within the scope of a patent. These steps have been in the art for decades, and the patentee did not invent them. The claim language is not sufficiently specific to disclose prior art. In addition, a hypothetical patentee does not claim the use of prior art that he did not invent, and therefore the patentee does not get the benefit of the novelty.

In the USA, Bilski v Kappos and the Prometheus Laboratories, Inc case are examples of the latest developments. These cases were filed in light of Bilski, and the Federal Circuit re-decided them on 17 December 2010. Then, in April 2011, the Supreme Court granted a petition for a writ of certiorari and ruled in favour of Prometheus.

Sui generis

A sui generis copyright is a multi-factoral contract that protects the use of an algorithm or digital therapeutic. Sui generis copyright can be obtained without applying. The protection can cover graphical interfaces, integrated multimedia elements, or even the look and feel of a telehealth application. Copyright does not protect the ideas and principles underpinning computer programs. Algorithms, functionality, data file formats, and graphic interfaces can all be protected.

The case in question involves the use of an algorithm to produce a downloadable database. The author argued that the CJEU had ruled that the database creator did not have the sui generis copyright. In a separate case, the CJEU ruled that the right holder did not have sui generis copyright over his database. However, the author emphasized the importance of posting general terms and conditions for users, as just posting the terms online does not mean they will be enforceable. Thus, the author argues that contracts are indispensable for controlling the use and access of publicly available data. This argument is particularly strong, since data exchanged under data access agreements can be controlled in a way that the right-holder does not.

While sui generis copyright is not appropriate in this case, it is the best starting point for computer software intellectual property protection. In addition to granting broad idea protection to certain elements of software products, it also provides a foundation for defining sui generis copyright. However, the application of sui generis copyright law is not appropriate in this case, because it would give no protection to the underlying idea, which may be the most important contribution of the developer. The failure to give such protection to the underlying idea can inhibit innovation. However, recent applications of copyright law have given expansive idea protection to certain elements of software products.