Unlike AI, algorithms already have a major impact on our lives today. This blog looks at the possibilities of legally protecting processes laid down in programs.
And the answer is: abstract and intellectual methods are not patentable. Mathematical methods that are behind or embody the code sequences are not protectable under patent law. Algorithms contain a process by means of which step-by-step operations are carried out. Such processes are not patentable.
Once the steps have been taken to see the arrangement of certain work steps in algorithms, copyright protection can arise either from the individual work steps. The same applies to these as to the copyright of software in general. The sequence in the source of the individual work steps is copyrightable, but not the idea behind the sequence. The reason for this is simply that the legislator is less interested in protecting the individual's "invention" than in protecting the national economy by ensuring that there can be no monopolies on ideas, processes, etc. § Section 69a (2) sentence 2 UrhG (Copyright Act) reads:
(2) ....S.2: Ideas and principles underlying an element of a computer program, including the ideas and principles underlying the interfaces, are not protected.
And the same applies analogously to algorithms. Ideas for carrying out certain work steps in a specific sequence are generally not eligible for protection.
Protection under copyright law: Possible, but difficult.
Of course, the arrangement of certain work steps can constitute an act of copyright. It is no different than in art. Picasso made sculptures from certain pieces of garbage and branches that he found on the beach. My favorite example is the baboon monkey with the child. The head is made from a model car. The parts themselves were already finished when Picasso found them. The arrangement of many parts that are not themselves protected by copyright can again be an act of copyright. But beware: the arrangement must not be so mandatory for technical reasons, because the lawyers would then conclude that the arrangement of the individual steps itself was not the result of a creative act, but the product of a mandatory "technical requirement". However, copyright law, which was created for artists and not for technicians, states that a creation can only arise where there is scope for design. And ultimately, one thing is also crucial: who is supposed to reveal their secrets and disclose the structure of the work steps in the context of an infringement suit without the other party benefiting from it? In the context of possible infringement proceedings, in order to prove that your own company's software has been "copied" by the other company, you would have to disclose your own software, with the risk of losing know-how.
So the only option left is to protect algorithms in accordance with the Trade Secrets Act. I have described elsewhere how this can be done. Some of the literature argues that algorithms that are protected under the Trade Secrets Act could be licensed on a contractual basis. I do not believe this. For decades, the Federal Court of Justice has resisted every attempt by the software industry (and in particular the attempts by the legal departments of Microsoft, IBM and Oracle) to include new absolute rights in the copyright canon. If the legislator makes the statement at one point that ideas should not be monopolized, this assessment cannot be counteracted with the vehicle of the GeschGehG.
And it's not that terrible. The software industry in Europe has been living with the "ideas and methods are not protectable" approach for more than 25 years and is thriving. Even if it is the job of lawyers to examine the possibilities of protecting intellectual property, it has to be said that start-ups are hampered by monopoly positions. And the majority of German software manufacturers have fewer than 50 employees and do not have the financial resources to employ lawyers to deal with intellectual property issues. Here, as in patent law, there is a point to some things.
