Jornal O Globo - June 12, 2024
Publications that challenge the control of platforms use everything from out-of-context material to deepfakes of famous people, which are increasingly difficult to detect.
The tactic of using AI-generated content has become increasingly common among extremist groups that spread hate messages online, challenge the control of platforms, and try to gain supporters with synthetic images, audios, and videos that spread violent speech. In a report of over 200 pages, a group of American researchers from the Middle East Media Research Institute (Memri) mapped out dozens of examples this year of how AI has become a weapon in the hands of these organizations.
— Extremist groups can create their own models and use them in different ways. In addition to the most direct, which is to generate texts and posts with content adapted to their speech, there are more innovative uses, such as creating weapon designs for 3D printers or even instructions for creating bombs — explains Fernando Ferreira, a researcher at NetLab at UFRJ.