Cesar Leite and Pedro Bocchese wrote a very interesting article on " Artificial Intelligence within reach of all"1addressing multiple connections of human beings with AI and highlighting the recent popularization of ChatGPT as "the first effective and functional connection of the vast majority of people with AI".
In this topic, it is important to highlight that for the existence of artificial intelligence a basic harmonious arrangement of a) organized data; b) algorithms; c) a platform with processing capacity is necessary. It is interesting to note that many software applications used in our daily lives already need and use this trinomial without this use characterizing them as artificial intelligence applications. But that is a simplified explanation of the bare minimum we need to understand who the "owners" of AI are.
When we talk about organized data, we have two possible specific protections, without going into other broader issues: 1) Copyright: in addition to the content of the data, if they characterize original texts/works, the "simple" organization of databases is also protected as copyright by Brazilian Law 9,610/98; 2) Trade secret: whether or not it has copyright protection, it is possible to protect this database organized by contractual and technological means so that it remains confidential/secret, being duly regulated by Brazilian Law 9,279/96.
Meanwhile as we go into the algorithms themselves, we also have two basic forms of protection in the area of intellectual property: 1) Trade secret (already explained); 2) When the algorithm is used to solve a technical problem in the context of describing an innovative process/method or inserted in a system, there is the possibility of protection through patents also covered by Brazilian Law 9,279/96 if they are fulfilled patentability requirements.
Finally, regarding the platforms that make it possible for algorithms in communication with organized data to generate an innovative technical effect with the speed demanded by the logic of their interaction, we also have two essential models: 1) patents (already explained but with the different application by the technological essence of the platforms); 2) Trade secret (already explained but with adaptations on how to keep confidential the know-how involved in "products/processes" that reach the market and may be susceptible to "reverse engineering" and/or quick overcoming by the learning curve and time/effort needed to inventing around.
The clear problems faced by new (and not-so-new) technologies are usually related to authorship, ownership, patentability criteria, and jurisdiction in which the technology protected by intellectual property is being disseminated. As an important element for the competitive advantage, the clear identification and protection of the intellectual property in the intellectual creations involving artificial intelligence made companies carry out a real technological race with an increase of more than 718% in the worldwide patent applications involving artificial intelligence and more than 699% increase in the number of patents involving Big Data technologies between 2015 and 2020, according to official data from the World Intellectual Property Organization - WIPO2.
But then, whose intellectual property is related to ChatGPT or DALL-E technologies? Within the technological races, the companies involved are usually looking to follow the public demarcations of property of their competitors in the world. Thinking about the above tripartite logic (organized data, algorithms, and platforms) it is not difficult to imagine that the race for the appropriation of artificial intelligence is not only going around in the West, but is also clearly on the rise (and, who knows, overcoming) in the East, as the example of Huawei developing an application similar to ChatGPT3.
It turns out that the open international systems of intellectual property are relatively efficient in publicly identifying the intangible assets protected by patents, trademarks, and industrial designs. However, they are not efficient enough to identify copyright assets corresponding to elements of artificial intelligence, software and, mainly, they are not efficient in publicly identifying intangible assets when related to trade secrets (which the name itself correspond to the difficulty of access to its content and gives meaning to the specific protection).
We reached to the conclusion that undeniably involves uncertainties regarding the use of the intellectual property as a competitive tool to enhance the business formats of technology holders involving artificial intelligence, as well as the risks of involuntary violation of third-party intellectual property rights by non-human creations made using artificial intelligence. As the creator of ChatGPT said in a recent interview for Time magazine: "artificial intelligence must be regulated".4.
"Article originally published on www.dynamicmindset.com.br on 28.02.2023."
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.