The impact on cultural diversity of Automated Anti-Piracy Systems as copyright enforcement mechanisms: an empirical study of YouTube’s Content ID digital fingerprinting technology

Investigators: Sabine Jacques (University of East Anglia), Krzysztof Garstka (University of Cambridge), Morten Hviid & John Street (both University of East Anglia)

The research aimed at understanding whether there is a correlation between the introduction of algorithms as copyright enforcement mechanisms and the diversity of works enjoyed by the wider public on online platforms. The main objective was to investigate whether automated technological tools foster or deter the promotion of cultural works in the digital economy.

The first part compares and evaluate the regulatory framework of key jurisdictions: the US and the EU with a focus on the UK. This enables the understanding of the birth and rise of these automated anti-piracy systems as copyright enforcement mechanisms. Furthermore, this part we explain why promoting diversity matters and how it can be achieved by ensuring that freedom of expression is preserved in the online environment. This includes the exercise of freedom of expression of all and not a handful of powerful actors.

The second part constitutes the empirical analysis. By looking at specific platform and a specific type of use, we sought to understand whether online expressions travel equally in a borderless environment and how algorithms as copyright enforcement tools influence the culture enjoyed in a particular society.

This research concludes that there is a correlation between the use of algorithms as copyright enforcement mechanisms and the promotion of cultural diversity in a particular society. Based on the current legal framework, platforms have strong incentives to avoid displaying content which might infringe copyright. The implementation of algorithms assists these intermediaries as the design of algorithm is likely to focus on avoiding the display of allegedly infringing content. Hence, there is a risk that these algorithms are over-inclusive. To put it differently, the balance struck by the legislator in copyright legislation is likely to be jeopardised in the online environment.

While part 1 establishes the direction of the effects of algorithms, the empirical analysis provides insight as to the magnitude of these effects. This research establishes an empirical analytical framework on how cultural diversity can be measured in the online environment. The preliminary results demonstrate that cultural expressions do not flow freely in a borderless environment and that there is a modest change in diversity due to the introduction of algorithms. Furthermore, there are important network effects operating on a platform.

Finally, this research demonstrates that promoting a better choice and access to cultural works online and across borders while establishing a fairer and sustainable environment for creators and creative industries requires considering the interplay between copyright, freedom of expression and diversity.

A draft version can be accessed at