To date, Artificial Intelligence appears to be a technology that can be adopted in many areas. A study by the European Parliament, in fact, estimated an 11%-37% increase in work productivity achievable by 2035 thanks to AI; moreover, exploiting this technology in public services could lead, by 2030, to a reduction of global greenhouse gas emissions equal to 1.5%-4%.
Therefore, Artificial Intelligence is destined to have a significant political and socio-economic impact. However, if it is true that AI can represent an opportunity, it is equally true that it is increasingly sophisticated, pervasive and – sometimes – non-transparent, to the point of being able to represent a risk for humans also in terms of protection of fundamental rights.
In recent months, there has been a lot of talk about ChatGPT, a conversational Artificial Intelligence tool, which has aroused some interest, but also raised some questions. It is a chat that allows you to reproduce natural language: when a user composes a message, ChatGPT processes an input that allows you to generate a response that is consistent with the content of the conversation. This is possible thanks to two aspects of the AI on which ChatGPT is based: machine learning, the so-called automatic learning according to which systems are able to learn from the data they detect, independently identify recurring patterns and make decisions by minimizing the human intervention; deep learning, based on algorithms that simulate the neural networks of the human brain and which allow deep or hierarchical learning.
Being thus characterized, ChatGPT brings with it possible problematic implications: the generation of fake news, understood as the dissemination of false information, but also as an incorrect classification of some contents deemed misleading, even if they are not such; the infringement of the privacy of any subjects involved; the configuration of crimes such as defamation – therefore damage to reputation – slander, dissemination of discriminating information, violation of copyright.
Artificial Intelligence can facilitate the creation of intellectual works, but it opens up some debates; specifically, ChatGPT has an international scope, which makes it difficult to regulate its use in a universally shared way, as each state has its own legal system.
“To clarify this aspect we must make a premise: according to Italian regulations, software and Artificial Intelligence have no legal personality and as such cannot be recognized as inventors” – comments the lawyer Vercellotti – “Also according to these regulations, in order for patentability or registrability to be established, the inventor must be identified. So, to whom should the originality of the work produced be attributed? And, consequently, to whom can the moral rights be attributed – ie the authorship which is inalienable – and the patrimonial rights of economic exploitation of the work itself? These are questions that – in my opinion – must necessarily be studied in depth, in order to arrive at legislation whose primary objective is the protection of man, especially in terms of fundamental rights”. conclude.
ChatGPTPlus was introduced in February, a paid version of Chat GPT that offers a number of advantages. “To date, not much is known, except that the use of the chatbot will be guaranteed even during peak hours, faster response times will be ensured, priority access to new features and improvements which, however, have not yet been specified by the ‘agency. From a legal point of view, therefore, we are waiting to see the real differences and their extent”.
Finally, a further debate remains open: it is clear how Artificial Intelligence manages to be a support for man, but is the fear that it could even replace him well founded?
“AI is capable of replacing humans in those repetitive tasks, such as transcription” – explains the lawyer Brunella Martino – “For example, tomorrow it could replace it in those transcription jobs, for which software would guarantee the maximum crystallization of information. My advice is to put quality into your work – which is what can distinguish our work from that of a machine – but it is also to be assisted by Artificial Intelligence: these tools undoubtedly have the ability to process a significant quantity of data, but they do not reach the intellectual capacity of man”.