Possibility of creation of IP works with the help of generative AI models

Since the adoption of the #DraftoftheAIAct in the European Parliament, there has been a lot of talk and opinions amongst lawyers (from Serbia as well) about whether something created with the help of the #generativeAI may be treated as intellectual property? And who is its titular owner?

Unsurprisingly, there have been a lot of lawyers with the opinion that this is not possible – in other words that what has been created with the help/assistance of #generativeAI may not be treated as intellectual property and may not be protected as such. Giving prompts is not enough to be the “artist” and create something unique that could represent intellectual property. These are very rigid and conservative opinions based on narrow interpretations of the currently existing legal framework. This reflects a profoundly deep misunderstanding of what #generativeAI is (or what it may be), which is, to be frank – not uncommon among lawyers. #GenerativeAI is simply put – a tool. Something to work with. Just like it was some 100-150 years the case with a photo camera. After all, it “only” captures something that is already existing out there in nature and the world, nothing was hand painted, was it? This especially is so when it comes to #generativeAI models which are meant for images and videos, like DALL-E 2, Midjourney, Stable Diffusion, etc…

The main point is that without a prompt it cannot produce anything. However, this is where we come to what is important – human spirit and skill, which is what is according to the legal standards necessary to produce something that can be treated as intellectual property. The point is that using any of these #generativeAI models like DALL-E or Midjourney does not disable people from creating something valuable that will be able to be protected as intellectual property.

Using tools like this demands at least two things: the first thing is imagination necessary to create something – to have an idea and the second thing is to have enough skill to provide adequate prompts to realize this idea one has. This is where the catch is – having skill and imagination. You need to first imagine something and then have adequate skills to transfer it into something in real life.

Eventually, using these tools will require skill and knowledge to create something specific and valuable you have previously envisioned in your mind. You will just be using an #AI model instead of a photo camera. It comes down to one question – if what you have created represent something truly original and new, something that did not exist before it was made by you? It is about having the skill to devise a series of very specific prompts to realize (supposedly) unique ideas you have and transfer them into media, which will in turn create output that will satisfy the legally prescribed criteria to be treated as intellectual property.

How will we know if this was enough to create something that can be treated as intellectual property?

Serbian Supreme Court of Cassation in one of its relatively recent decisions took the position in one of its decisions that there is no originality in works that are the result of conscious or unconscious imitation or imitation of already existing works, which distinguishes it from other works.

This provides a glimpse of reasoning that may give way to the recognition of IP rights for works created with the assistance of #generativeAI. If it satisfies the prescribed criteria given in jurisprudence, why wouldn’t it be recognized as IP? Only because it was created with the assistance of #generativeAI? This is what painters thought of early photographers and still, here we are – photos may be the subject of IP. We have litigations over who is the author of these photos. Naturally, if it meets the prescribed criteria.

It is all about the human contribution and idea behind output which was created with the assistance of #generativeAI.

Ultimately (as always), the law and jurisprudence and precedents will give way and will have to accommodate the new reality. Laws will change to follow and reflect life, as they always do when there is a discovery before it becomes an essential and common part of our lives.

So, to answer the question from the beginning – yes, the author of this article considers that the outputs created with the assistance of #generativeAI may be recognized as intellectual property and that it will only be a matter of satisfying certain legal standards in the process (which will naturally evolve). The owner will be a human person who put in enough effort and skills to realize their unique idea.

What we will have to be careful about is the ethics of it all.