The new Radar Tecnológico, produced by the ANPD’s Technological Innovation Bureau (SITEC), stands out for focusing specifically on the phenomenon of deepfakes. In the study, deepfakes are defined as audio, images, and videos created or altered by artificial intelligence with a high degree of realism. The Radar also addresses the conceptual boundaries between deepfakes, cheapfakes, and legitimate uses of synthetic media, including in sectors such as audiovisual production, healthcare, and education.
A central point of the study is the connection between deepfakes and the General Data Protection Law (LGPD), since manipulating faces and voices draws on multiple categories of personal data. The publication also devotes a section to the Brazilian context, listing among the main risks non-consensual pornography against women and girls, digital racism, electoral disinformation, gender-based political violence, and financial scams.
The document makes clear, right in its introduction, that it is technical-informational and forward-looking in nature, constituting neither regulatory guidance nor a decision by the Agency on specific cases. Even so, it is published in the context of Decrees No. 12,975 and No. 12,976, of May 2026. The first amended the regulation of the Internet Civil Framework (Marco Civil da Internet), establishing for providers duties of care, systemic risk management, and greater transparency, while the second set out guidelines for combating violence against women online, with a chapter dedicated to the generation of intimate content by artificial intelligence.
In its final chapter, the Radar Tecnológico devotes a specific section to the “legal and regulatory perspective,” in which it maps three broad patterns adopted internationally: transparency and labeling obligations for synthetic content (exemplified by Article 50 of the European AI Act), prohibitions on use in electoral contexts, and the criminalization of abuses. Even so, the study’s focus is not primarily regulatory. For the ANPD, deepfakes are part of a broader transformation in how society produces, circulates, and validates information, and addressing them will depend on technical solutions and specific legal instruments, but above all on strengthening digital literacy, media education, and algorithmic transparency. It is within this framing that the Radar situates the regulatory discussion: without closing it, it shows that the topic continues to evolve and remains on the Agency’s agenda, potentially informing future initiatives, should they prove necessary.