The results that generative artificial intelligence systems are showing and the risks they entail are worrying and more and more people are wondering if there is a way to protect themselves, if there are legal instruments to defend themselves against AI or how to do it.
“Soon we will have two European regulations (the IA Act and the Cyber Resilience Act) that will specifically regulate AI and that will introduce prohibitions, obligations and sanctions; but in the meantime it is not that we are helpless, because we have well-defined regulations on data protection, copyright and the right to one's own image that we can apply”, says Albert Agustinoy, coordinating partner of the intellectual property group of the Cuatrecasas law firm.
And it details that there are several precepts of the data protection regulations that are directly required of systems such as ChatGPT: transparency regarding the use and treatment of the information that it collects from the user, the obligation to inform that it is an automated treatment and the need to offer the option of being supervised by a human, or the obligation of the company that wants to use an AI to process personal data to carry out an impact assessment on the rights of individuals.
As generative AI feeds on pre-existing information, Agustinoy recalls that it can also be prosecuted via copyright infringement, as Getty Images has already done before the High Court of Justice in London. And to fight against deepfake, against the use of someone's image or voice to falsify dialogues or scenes through AI, "the legal framework on the right to one's own image can be used."
"It is true that the law lags behind technological advances, but we can apply the regulations that we already have by adopting criteria to adapt it," says the Cuatrecasas expert.
And he stresses that it is also the job of companies and organizations to monitor the professionalized use made of these technologies and instruct their employees. "At the firm we have established guidelines prohibiting the use of ChatGPT in relation to legal advice because we consider that it is not yet reliable and that it entails the disclosure of confidential information", he exemplifies.
All in all, Agustinoy believes that, in view of the reactions provoked by ChatGPT, the European organizations will speed up the procedures for the two new regulations that it has in place, the one on AI and the one on cybersecurity, so that they are ready this year.
Less optimistic about the deadline are the CSIC AI researcher Ramon López de Mántaras or the mathematician and astrophysicist Manel Sanromà, one of the promoters of CIVICAi, the first citizen lobby to defend human interests against AI.
“The unit of measure for AI progress is the week, and bureaucrats are not used to that pace; For this reason, the first thing is that governments assume that we are in a critical situation that requires acting now, not launching an AI Agency in 2024 because in a year we do not know where this technology will have reached ”, says Sanromà.
He adds that, in addition, regulations and sanctions are necessary but not sufficient to control AI. “It is necessary to immediately implement a strong AI agency in which scientists, companies, governments and also civil society participate, and with the capacity to monitor what is being done in laboratories, as the International Organization does with nuclear energy. of Atomic Energy”, says Sanromà.
He assures that, more than sanctions, what is needed to control AI is "an agency with a lot of scientific capacity but also with a citizen presence that guarantees transparency and the defense of people's rights."
