Summary

The report first gives a brief introduction on the topic of deepfakes, after which it looks at how current Dutch law responds to various deepfake situations with its substantive law, procedural law and enforcement rules. It also devotes some attention to some external legal examples concerning the (attempted) regulation of deepfakes and provides some general reflections on the topic. The most interesting parts of the report are its main findings and the possible regulatory options that it introduces to meet the challenges identified in the report.

Main finding of the report: legislation is not the problem, but the enforcement is

The main finding that comes foreward within the report constitutes the fact that the legal framework applicable to deepfakes is not the primary problem; the problem is the enforcement of the existing and any future legal rules. Although legislative amendments are possible and perhaps even desirable in some cases, these would not tackle the main problem with regard to deepfakes in horizontal relationships. Committing fraud and deception by means of a deepfake is already prohibited, introducing false evidence in a court case is already prohibited etc. The GDPR also installed several obstacles to the creation and distribution of deepfakes. For example, the duty to inform portrayed persons that they are featured in a deepfake. The core problem lies with the enforcement: this is made especially difficult because of the fact that a huge amount of manipulated material has already appeared. Experts expect that this will only increase in the future and to complicate matters even further, deepfakes will also become increasingly difficult to detect.

Regulatory options

To relieve enforcement pressure, among other things, the researchers describe a wide range of regulatory options. However, a number of caveats are introduced together with these regulatory options. The report firstly stresses that regulatory options are not recommendations, but options; their desirability and feasibility will have to be subject to further research and political/societal discussion. Secondly, some of the options can be implemented immediately, while others require structural changes to the legal system, and still other options are controversial or have major potential negative consequences and require further research. Moreover, the regulatory options should be considered in relation to each other. Several options address the same underlying problem, meaning multiple options could be introduced and can be considered complementary but introducing one of them may also suffice. Lastly, many of the problems described and potential solutions offered are related to general, societal trends. Sometimes it is possible to adopt specific rules for deepfakes; often, however, it seems advisable to address the underlying problem as such.

One interesting proposal put forward by the report is to prohibit the production, offering, use and possession of deepfake technology for the consumer market altogether. Current legislative provisions do not restrict the development of deepfake technology and its offering, but only its use for specific purposes. The background to this proposal is the report's finding that more than 95% of the deepfakes currently published between citizens are potentially punishable.

The report proposes a convenient scheme of different regulatory options spread across several domains, which can be download below.