The emergence of ChatGPT has begun to raise severe legal doubts about its use. And so severe that large financial institutions in the United States and Europe – such as Bank of America, Citigroup, Deutsche Bank, Goldman Sachs, Wells Fargo and JPMorgan – have decided to stop using the technology as a precautionary measure.
“It´s dangerous to use this platform because data protection is not guaranteed when consulting on a confidential case. Moreover, Open IA reserves the right to use the data provided to this application,” says Alejandro Payá, a partner at Cuatrecasas, a law firm that has also decided to prohibit its use by employees.
The way this platform works is very different from other search applications such as Google, explains Eudald Camprubí, founder of the startup Nuclia, which operates in this sector. “ChatGPT and its counterpart platforms are considered companies that operate as APIs (Application Programming Interface). In order to give the right answer, these companies need the information to travel to their computer system. In this way, confidential information leaves their infrastructure and the customer loses control of his or her private data. This is very different from using Google. This search engine takes the searcher directly to a third party website without the need to process the search to give an elaborated answer”.
Beyond the protection of confidential data, which is not guaranteed at the moment, the use of ChatGPT and its counterparts may raise questions about the authorship of texts and their dissemination. Who should appropriate a text or an image generated by the platform? For the time being, notes Manuel Martínez, a partner at Across Legal, a law firm specialising in technology, it is not necessary to mention that the text has been produced by the robot, as its level of creativity remains low. “Despite the fact that automated language is becoming more sophisticated, the algorithm’s response is still basic and not comparable to that of a human. Without the element of originality, there is no copyright and therefore no need to mention that the text is your work. However, the lawyer considers that for ethical or moral reasons it would be desirable to mention the robot’s authorship. Ignacio González, founding partner of the law firm dPG Legal, adds that in any case the question depends on the agreement between the provider of the tool and his client.
Going a step further, Martinez believes that the algorithm does violate copyright when creating content. “The generation of images and text is carried out from small pieces of pre-existing photos or content,” he argues. González qualifies this position. “I don’t see a problem if the robot creates the text through open databases that do not infringe on third-party creations,” he argues.
If that were the case, that would lead to another problem: determining who is guilty of copyright infringement: the robot itself or the person behind the machine? There seems to be no doubt: “Current law does not yet attribute legal personality to the robot. In that case, the person responsible would be the programmer or the chain of programmers who developed the algorithm”, says Martínez, who admits that it may be difficult to find the culprit in a sector as globalised as the technology sector. The culprit, says González, could be fined for plagiarism, distribution or unauthorised public communication and face fines of between 150,000 and 600,000 euros. In addition, companies that exploit technology could also be liable for other crimes such as disseminating hate speech or violating the right to honour.
Source: La Vanguardia
