#Interviews
#Interviews
04.03.2026
Only 12 Brazilian universities regulate the use of AI
USP survey finds that artificial intelligence governance is progressing slowly in higher education; expert calls for a pedagogical, nonpunitive approach
AI regulation in Brazilian universities is still in its early stages, but discussions about the ethical use of technology are gaining momentum across academic institutions | Image generated by AI
With the rapid expansion and consolidation of artificial intelligence (AI) services, universities and higher education institutions around the world have begun developing guidelines for their use. In Brazil, however, the number of institutions with formal regulations remains relatively small.
According to a survey conducted by the research group “Governance of AI Agents,” affiliated with the Institute for Advanced Studies at the University of São Paulo (IEA-USP), only 12 Brazilian institutions have defined and published guidelines governing the use of AI by students, professors, and researchers.
Journalist Jonas Gonçalves is a member of this group, led by researcher Virgílio Almeida. Prior to joining the initiative, Gonçalves researched the use of AI in journalism for his doctoral dissertation in media and technology at São Paulo State University (UNESP).
Within the USP group, whose primary focus is the study of AI regulatory frameworks, Gonçalves coordinates a subproject dedicated to examining the use of AI in higher education.
The objective is to generate data on the integration of AI in Brazilian universities and to contribute to the development of governance models that can serve as references for other institutions implementing the technology.
Science Arena spoke with Gonçalves to discuss his perspectives on AI regulation in Brazilian higher education.
Science Arena – Only a few institutions in Brazil have regulated the use of artificial intelligence. Why is there still resistance to incorporating this technology into higher education?
Jonas Gonçalves – To understand precisely what is happening at each institution, we would need to conduct individual assessments. However, we have identified some possible factors that help explain the current situation.
For example, each institution has its own layers of bureaucracy, involving multiple decision-making bodies. In such contexts, processes can move slowly. On the other hand, as we have observed in some cases in Brazil, progress can be faster when there is already an organizational culture for such.
A good example is the Federal University of Minas Gerais (UFMG). The institution created a committee to discuss the use of AI within its academic community.
From these discussions, UFMG developed a set of recommendations for the use of AI in the university context, guided by principles designed to promote understanding on how to use AI at the university.
Is the model adopted at UFMG a good example of AI regulation?
I believe it is one of the best approaches available. In our research group, we advocate for a pedagogical model of AI governance—one that favors understanding over outright prohibition.
The more institutions invest in training students and faculty to understand how these systems work, as well as the ethical boundaries involved, the more effective the integration of AI into Brazilian higher education will be.
Under this governance model, the tendency is to adopt a more guidance-oriented and conscious approach, which encourages transparency. If the prevailing logic is punitive, however, people may feel reluctant to disclose their use of AI systems. As a result, they may continue using these technologies without openly acknowledging the fact.
This dynamic can increase the likelihood of ethical issues related to AI.
But what are the ethical boundaries when it comes to this technology?
Universities already had general ethical principles in place before the emergence of AI. The challenge now is that with these artificial systems, there is a need for specific ethical principles, because there are particular routines that require limits on the use of these technologies.
One example is the use of similarity-detection tools, which identify passages in a text that resemble previously published material. However, textual similarity does not necessarily imply plagiarism.
Plagiarism itself was already addressed in university regulations prior to AI, but now the big difference is in the scale.
From the moment automated tools exist, the ability to produce content increases exponentially. Therefore, there is concern about maintaining control over this.
Curbing ethical infractions such as plagiarism is absolutely correct. However, we seek to emphasize the need to differentiate between the conscious, ethical, and transparent use of AI and the misuse of the technology.
Critics of AI use in universities argue that the technology may have negative effects, such as a decline in intellectual capacity. What is your opinion?
I believe this is a very valid discussion, but it is still too early to draw definitive conclusions. There is, in fact, a tendency to delegate tasks to AI. However, it is important to stress that the technology can at most serve as an assistant, for example, by helping with data collection. Under no circumstances should AI be considered a coauthor of a work.
Although AI offers certain conveniences, human intellectual work must continue to exist. This is already a consensus in the guidelines implemented for AI use.
How do you view the discussion of AI regulation in educational institutions when comparing Brazil to other countries?
From what I have observed, discussions on this topic do exist in Brazil. However, universities are also proceeding with caution.
AI evolves very rapidly. This means that a guideline implemented at one moment may not be sufficient for future situations. This creates some apprehension among Brazilian institutions.
Compared with other countries, I do not believe Brazil is lagging behind. According to the Ministry of Education’s Higher Education Census, there are 2,561 active higher education institutions in the country.
Of that total, we identified twelve institutions with published guidelines on the use of AI, according to our mapping.
It is a very small number, but that does not mean that only these 12 institutions are discussing the issue or have established rules. Other institutions may already have regulations in place, though limited to specific departments or courses. I believe this topic is making progress in Brazil, which is already a positive sign.
*
This article may be republished online under the CC-BY-NC-ND Creative Commons license.
The text must not be edited and the author(s) and source (Science Arena) must be credited.
