In recent times, AI and ‘chatbots’ (AI-based computer programs capable of maintaining a real-time conversation via text or voice), such as ChatGPT, have entered the common lexicon, posing threats to the production of original texts by an author and increasing the risk of plagiarism.
However, more than the texts, the creation of visual artistic works begins to raise real concern, namely through the DALL-E program, from the company OpenAI (the same company that developed ChatGPT), in which it is not possible to find the author in the origin, told Lusa the lawyer Manuel Lopes Rocha, specialist in Intellectual Property Law and Technology.
“There are two phases to the recent discussion about copyright protection for authors. Copyright is very anthropocentric. What jurisprudence is saying around the world right now, and on AI issues, is that there always has to be a human behind the creations”, he explained.
This worldwide discussion has been driven by the British Stephen Thaler, who has tried to patent a work entirely made by an AI system, the DABUS (‘Device for Autonomous Bootstrapping of Unified Sentience’), which he owns.
Stephen Thaler has tried to change the position of the courts and registrars, but has so far lost patent litigation everywhere except South Africa, which has recognized DABUS as an inventor.
“But now we have entered a new phase: until now it was possible to find a human being at the base of all this. At the moment, with this software (DALL-E), by OpenAI, the problem is that it is no longer possible to find , almost, the human in origin”.
This is posing problems for a copyright that has always been, since its inception, anthropocentric.
“If the use of the ‘chatbot’ databases gives rise to an infringement of a published work by an author, then there will be a violation of copyright. In the case of painting, in which this ‘software’, DALL-E, who makes works where it is very difficult to find the initial author, there are copyright problems, dogmatic problems and this is an extremely interesting question”, considered the jurist.
At a time when we are walking on unknown ground, we are experimenting with some variations in the area of copyright: in the United States, a work containing images, the so-called “generative AI”, was registered, in which the Copyright Office of the country accepted the registration of this work, saying that these images were generated, despite everything, under author control, “but (it will) reach a point where it is difficult to find an author, even a remote one”.
Manuel Lopes Rocha recalled that in 1988 there was an amendment to the English intellectual property law, even before the ‘software’ directive, which inserted a rule, according to which the copyright can also belong to those who arranged the means to produce the work .
“At the time, that standard was heavily criticized, it was considered completely futuristic, something (in the style of science fiction author) Philip K. Dick. What seems to me is that in areas such as AI ‘software’, in the area of painting, it is posing problems for us, because the initial author may no longer be found. It is very complicated and this actually poses problems for the law traditional author”.
The lawyer recognizes that it is necessary to start acting to fill this legal void, but says that legislation on the matter is not yet being legislated, because “this is something very surprising” and “a large part of the discussion comes from November of last year”.
What is happening right now is that OpenAI seeks to create “a system, a kind of ethical code, in which those works that are completely created by AI will be identified”.
On the other hand, “the European Union is currently studying legislation on AI, but this is great news, because for the first time the human author is at stake, without a doubt”.
Also Read: Dropbox announces layoffs of 16% of workers
