Beviskraft och EncroChat - Felkällor vid bevisvärdering

Sammanfattning: After the French police managed to gain access to material from the encrypted communication service EncroChat, parts of the material were forwarded to the Swedish police. This resulted in a large number of people being sentenced to long prison terms. EncroChat is partly an encrypted communication service and partly an operating system that runs on smartphones. EncroChat is used by people who do not want the content of their communication to be disclosed and/or linked to them personally. On EncroChat users oftentimes communicate through randomly generated usernames, consisting of two English words. The Swedish National Forensic Center (NFC) has compiled information on the service and its functions, but the information is not exhaustive and ambiguities remain. Several courts have concluded that parts of the material, which was received by the Swedish police, are missing. What this is due to is still unknown. In addition, the material has also been processed by Swedish police before it is presented to the court. As there is still a great deal of uncertainty surrounding the EncroChat material, inter alia due to the fact that the service itself is a relatively new phenomenon, this places high demands on the courts regarding how the material is evaluated. Several courts have stated that this type of evidence should be evaluated with caution. Despite this, there are several examples of convictions based (almost) exclusively on EncroChat evidence. What a cautious evaluation of the evidence in practice means hasn’t been developed to any great extent by the courts. To carry out a safer and more accurate evaluation of evidence, probability theory can be used as a tool. Bayesian probability theory is well suited for this purpose, as it clarifies mistakes that frequently occur in the process of evaluating evidence. Bayesian probability theory forces the evaluator to structure their evaluation and to take a position on certain issues that are decisive for it. The thesis presents and explains central parts of the theory to show how one can use it to calculate the probability of a hypothesis in a structured and relatively safe way. This includes common errors in the evaluation of evidence, which can occur regardless of whether the evaluation is done with or without the help of Bayesian probability theory. The thesis also provides potential psychological explanatory models for these types of mistakes. The thesis presents the Swedish courts' evaluation of EncroChat evidence in seven different verdicts, of which four decisions derive from the district court and three decisions derive from the Court of Appeal. The courts' evaluation of evidence is then analyzed based on Bayesian probability theory to determine whether the evaluation of evidence contains common sources of error, and whether it is consistent and uniform among courts. The thesis concludes that there are strong indications of sources of error in the courts' evaluation of evidence. The evaluation of evidence does not seem to be consistent and uniform among courts.A possible reason for this is the lack of a tool to systematize the evaluation of evidence in general and to avoid sources of error in particular. An evaluation of evidence based on Bayesian probability theory could reduce the frequency with which misjudgments occur and lead to a more consistent and uniform evaluation of evidence among courts.