The ethical issues surrounding deepfakes stem from their potential to exacerbate injustice and harm. At the individual level, deepfakes can result in wrongful convictions or acquittals, violating principles of fairness and due process. Innocent individuals may be falsely implicated through fabricated media, while guilty parties may exploit the "liar’s dividend" to dismiss incriminating evidence (Chesney & Citron, 2019). Such outcomes erode the ethical foundation of the legal system, which is grounded in truth and justice. Increased AI systems play into the equity dynamics of deepfakes. Trained data sets are not always diverse enough for systems to accurately discern manipulations involving marginalized persons. According to Maras and Alexandrou (2019), detection tools that are flawed will yield inconsistent results across races or genders, compounding inequities established through the structure of the criminal justice system. Here, ethics surround regulations of fairness, responsibility, and accountability of AI developments when used as evidence toward establishing guilt or innocence. In addition to collateral damage comes the ethical framing of harm based upon what is seen. If a deepfake falsely impliculates someone in a crime or staged violence, the reputation of the afflicted party will suffer psychological and social harm that cannot be easily reversed. Even if one is found not guilty due to such a deepfake, the stigma of being caught on film for such a crime lingers. According to Westerlund (2019), this is especially problematic since what happens on the internet stays on the internet.