Trust in User-generated Evidence: Analysing the Impact of Deepfakes on Accountability Processes for Human Rights Violations (TRUE)

用户生成的证据--定义为由普通公民记录并用于法律裁决的信息--在问责过程中发挥着越来越重要的作用。在世界各地，移动电话技术的进步和互联网接入的增加意味着数以百万计描述大规模侵犯人权行为的重要照片和视频已经并将继续在网上创建和分享。瑞典、德国、荷兰和国际刑事法院等国的大规模暴行审判已经利用了这种证据，联合国人权理事会授权的调查委员会、实况调查团和调查委员会也是如此。然而，与此同时，公众越来越多地面临着深度假冒的例子--使用机器学习技术创建的超逼真图像、视频或音频记录--随着技术的进步，它们只可能变得更先进，更难检测。这两个发展构成了一个重要的难题：对深度假货的看法是否导致了对用户生成证据的不信任？如果是这样的话，这对这种证据在未来人权问责进程中的作用意味着什么？到目前为止，许多文献都表达了一种担忧，即深度伪造的增加将导致对用户生成的证据的大规模不信任，这反过来将降低其在法律程序中的认识价值。情况很可能是这样，但目前还没有研究检验这一假设。这是一个亟待解决的重大证据缺口。通过在法律、心理学和语言学的交叉领域采用创新的跨学科方法，这一开创性项目将在使用非人道权利问责程序的具体背景下，首次系统地说明对用户产生的证据的信任。

User-generated evidence - defined as information recorded by an ordinary citizen and used in legal adjudication - plays anincreasingly important role in accountability processes. Across the world, advances in mobile phone technology and increasinginternet access mean that millions of important photographs and videos depicting mass human rights violations have been, and willcontinue to be, created and shared online. Mass atrocity trials in Sweden, Germany, The Netherlands, and the International CriminalCourt, amongst others, have already utilised this kind of evidence, as have UN Human Rights Council-mandated commissions ofinquiry, fact-finding missions, and investigations. Yet, at the same time, the public is increasingly confronted with examples ofdeepfakes - hyper-realistic images, videos, or audio recordings created using machine learning technology - which are only likely tobecome more advanced and difficult to detect as the technology progresses. These two developments pose an importantconundrum: have perceptions of deepfakes led to a mistrust in user-generated evidence? And if so, what does that mean for the roleof such evidence in future human rights accountability processes? Much of the literature to date has expressed a concern that the risein deepfakes will lead to mass mistrust in user-generated evidence, and that this in turn will decrease its epistemic value in legalproceedings. This may well be the case, but no study has yet tested that assumption. This is a major evidence gap that urgently needsto be addressed. Through an innovative interdisciplinary methodology at the intersection of law, psychology, and linguistics, thispioneering project will develop the first systematic account of trust in user-generated evidence, in the specific context of its use inhuman rights accountability processes.