The recent rise of artificial intelligence (AI) solutions has also had a significant impact on electoral processes. Most notably, deepfakes created by generative AI applications can (and have been) used to spread misinformation during the campaigns, but they can also be used for cyberattack automation, biased social media bots, etc. We instead present a positive use case for generative AI in manipulating video material required as proof of voting by potential coercers. For this, we have created a pipeline that takes a video of a voting ballot and replaces its critical content (in our case, the digits on the ballot). To achieve this, a YOLOv11 model is used to find the digits, a WavePaint image inpainting model is used to cover up the old digits and a separate image of the new digits is used to place them into the video. Additionally, we implemented the prototype application in the form of a web service and validated the outcome by asking the humans to distinguish fake ballot images and videos from the real ones. Our results show that humans can still recognise fake ballot videos relatively well, implying that requesting ballot videos is still an efficient attack against voter freedom in the case of paper voting. However, future developments in the generative AI techniques are likely to improve the situation significantly.
Habanen et al. (Thu,) studied this question.