Ethical and societal challenges in the use of generative AI in communication

This paper presents a narrative review describing the challenges of adopting artificial intelligence tools for audiovisual communication, identifying issues related to the ethics, control, transparency and legal framework. We make a critique of the term and delimit the research to generative neural-network-based computational processes. The problems are examined in three proposed areas: the input, which includes the training material; the process, which deals with the systems necessary to run these services, their development and control; and the output, in which the use of generated artifacts is analysed. International legal and judiciary approaches are reviewed in each step, as they are fundamental for the development of the study. Finally, some hypothetical future scenarios based on current AI debates are imagined, which help readers foresee possible situations regarding the use of generative systems. We conclude that although it is unstoppable that these technologies are adopted, their use might have some limitations, regulations or conditions. More importantly, the general use of AI tools will bring a shift in the way we produce and consume information. While we can clearly foresee dramatic increases in productivity, it is not clear how this production will be accepted by society. On one hand, studies have already revealed that subjects rank AI-generated content higher than content made by humans. On the other hand, the plethora of artificial media might provoke an overall rejection of the digital world.