When AI talks politics: a critical look at Deontology and Utilitarianism in the age of synthetic campaign content

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Until very recently, when someone wanted some fabricated political content that appeared authentic, they needed a professional production team. Nowadays, anyone owning a computer is capable of doing it in merely several minutes. Artificial Intelligence makes people able to generate fake speeches. Thereby, this technology enables the generation of fake videos together with customized campaign ads appearing to be true ones. This raises an important ethical question for media ethics. This paper compares two well-known ethical theories to find out their respective views on such topic. They are Immanuel Kant’s deontology claiming people are obliged by virtue itself to follow certain moral imperatives no matter what situation people find oneself in (Kant, 1785). Another one is Utilitarianism proposed by John Stuart Mill which evaluates any deed according whether outcomes make most people feel satisfied and happy (Mill, 1861/ 2002). The paper takes up one actual example from the elections happened in Malaysia in April 2024 to find out how the above two kinds of ethical doctrines deal with that issue in reality. Someone utilized AI technology to reproduce the voice of one candidate, so everyone thought that the one had completely abandoned rural voters. As shown by the paper, there are different results when adopting deontology versus utilitarianism respectively. According to deontology, the individual has acted wrongly in any event no doubt about it. But for the viewpoint held by utilitarianism, it seems most probably he got something wrong but perhaps his behavior would gain legitimacy under some other conditions. Neither of them seems fully convincing alone when judged individually. Therefore, the conclusion presented by the paper here is a compromise between the two extremes called here Hybrid Framework which sticks to deontological principles but still admits utilitarian exceptions in case things happen in really serious cases. Finally, the conclusion section will give guidance and advice on relevant issues raised before to everyone in different walks including ordinary people, content creators and policymakers.