This thesis resolves around problems arising for the existing legal framework from the use of novel software-­‐based policing tools during criminal investigations. The increasing dependence on information and communication technologies and the Internet means that more aspects of people’s lives move online, and crime follows them. This has triggered the development of innovative, autonomous investigative technologies that are increasingly replacing human officers for the policing of the online sphere. While only recently discussions of the legal status of embodied and unembodied robotical devices have gained more widespread attention, discussions of the legal status of autonomous agent technology are not new. They have focussed however in the past on applications in the private domain, enabling contract formation online. No systematic study has so far been carried out that looks at the use of autonomous agent technology when deployed by state actors, to fulfil core state functions. This thesis starts with the hypothesis that the use of automated, intelligent devices to replicate core police functions in the online world will increase in the future. Looking at first emerging technologies, but with an eye towards future deployment of much more capable software tools that fulfil policing functions on the Internet, this thesis looks at the challenges this poses for regulators and software developers. Based on extensive qualitative research interviews with stakeholders from two different jurisdictions (Germany & UK) this thesis finds that these novel policing technologies challenge existing legal frameworks, which are still premised on the parameters of the offline world. It therefore develops an alternative governance model for these policing tools, which enables their law-­‐compliant use and prevents rights violations of suspects. In doing so it draws upon both worlds, the technical and the legal, while also incorporating the empirical research results from the interviews with experts. The first part of this thesis analyses the technical foundations of these software-­‐based policing tools. Here, one of the key findings is that the current governance system focuses on ex-­‐ante authorisation of very specific, individual software tools without developing a systematic classification. This contradicts the principle of sustainable law making. To overcome this piecemeal approach, as a first contribution to existing research this work defines a new class of investigative technologies – mobile, intelligent and autonomous (MIA) policing tools ‐ based on the findings of the technical analysis. Identifying such a natural class of present and future technologies that pose the same type of legal issues should facilitate the sustainable governance of these new policing tools. The second part of this thesis analyses two specific legal issues: cross-­jurisdictional investigations and the evidentiary value of the seized data. These issues were identified as most pressing by the experts interviewed for this work. This analysis reveals that investigative activities of MIA tools are potentially in conflict with international law principles and criminal procedure law. In order to gain legitimacy, these new policing tools need to operate within the parameters of the existing legal framework. This thesis argues that given the unique technical capabilities of MIA tools, the primary approach to achieving this is to assign legal responsibility to these tools. The third part of this thesis develops a novel governance approach to ensure that MIA tools operate within the parameters of the legal framework, and therefore obtain legitimacy and relevance, also with regard to the investigative results. This approach builds on existing research identifying code as a regulatory modality and contributes to the field of legal theory. It constitutes a solution for the governance problems of MIA tools, however, it requires currently lacking collaboration among stakeholders and cross-­disciplinary research.