We would like to thank Valiente Fernández et al. for taking our article1 published in the previous edition into consideration. Our manuscript concluded with the following sentence: "The authors believe it is necessary to generate discussion on this matter within our scientific community." With their response, they are indeed opening and fueling a debate that seems more pertinent and necessary than ever. As Fernández et al. say, the creation of regulatory frameworks without follow-up discussions can lead to significant biases.
We, the authors, also do not believe that artificial intelligence (A.I.) should be excluded from scientific production, if such a thing is even possible. Since the publication of our ideas last month, the number of A.I. applications in the field of medicine has continued to grow and improved rapidly. The volume of tools that have been created is so significant that "directories" have been arranged to group all these tools together.2
The authors discussing our article correctly say that artificial intelligence is not inherently generative. We agree with this statement. However, we must always remember that A.I. is generating false data in some cases amidst a set of real data, something that is still under discussion.
It has also been noted how specific regulatory frameworks could quickly become obsolete in this rapidly evolving field. This is an indisputable fact, not only in this domain but in many others as well. However, unlike the world of laws, where legal frameworks tend to trail society, scientific communities have the ability to adapt constantly and evolve based on the variables on which they stand strong. Therefore, and without ever forgetting the rapid mutation of this topic, frameworks need to be constantly reviewed to avoid obsolescence.
In line with Fernández et al., we agree that applying regulatory frameworks is tremendously challenging due to the theoretical, technical, and entrepreneural complexity of A.I. However, we also see it as almost unavoidable that multidisciplinary teams should address this issue. This is no different from what working groups from different medical specialty societies do on a regular basis when they are working as a team.
We understand that initiating regulatory steps at the level of the entire scientific production can be a titanic endeavor. However, our focus on the drafting of scientific articles can be an important first step. And this does not have to be a complex or lengthy task to undertake. Some publishing groups already include in their guidelines some words dedicated to the use of A.I. regarding authorship, for example.3
As our colleagues have said, opposing technologies like these would be like opposing the very invention of the printing press. However, let us remember that even the printing presses that publish our research on a routine daily basis have their rules and regulations even though they don’t write papers themselves.