Improving Right Holders’ and Platforms’ Incentives to Design Automated Filtering

Various measures to regulate automated filtering have been discussed in recent years. However, only few measures have been evaluated in terms of efficiency. This chapter conducts the first comprehensive analysis of how right holders’ and platforms’ incentives can best be steered towards efficiency. For this, it identifies inefficiencies of right holders’ and platforms’ incentives in the absence of regulation and assesses to what extent and at what costs the discussed measures mitigate these inefficiencies. The analysis shows that without regulation, right holders have excessive incentives to prevent copyright-infringing uploads and deficient incentives to prevent false positives. It suggests negligence-based liability of right holders for false positives caused by deficiencies of the reference database. The analysis further demonstrates that without regulation, platforms invest too little into filtering and do not take some desirable accompanying measures. It suggests negligence-based liability of platforms for errors caused by the analysis of upload attempts. Platforms escape liability by complying with minimal requirements, which are adjusted as new learnings are collected. Platforms are further required to involve representatives of stakeholders in the development of filtering policies and to take the desirable accompanying measures. Based on these results, shortcomings of current US and EU law are identified.