Datafication-enabled advertising and other datafication practices, in the absence of proper constraints, will deepen the perils of datafication. A set of cross-border competition disciplines proposed in Chapter 5 may well be an effective instrument to address problems associated with platform monopolies and data capitalism. In this way, there would be less need for ex-post competition law enforcement in developing countries and LDCs, where relatively limited resources can be allocated to combat digital cartels and data monopolization. In this context, algorithmic transparency can serve as a starting point for global platform governance. The case study in Chapter 5 investigates the key dimensions of platform transparency requirements in a comparative context and demonstrates that the fragmentation of platform regulation is growing. The proliferation of platform regulations and algorithmic disciplines may place SMEs in an even more difficult situation vis-à-vis big tech companies, which have the resources necessary to manage different legal requirements in different countries. Despite the inherent complexity of the political economy surrounding digital capitalism, Chapter 5 concludes that there are reasons to be optimistic about better governance through international trade agreements.

Transparency has been in the crosshairs of recent writing about accountable algorithms. Its critics argue that releasing data can be harmful, and releasing source code won’t be useful.1 They claim individualized explanations of artificial intelligence (AI) decisions don’t empower people, and instead distract from more effective ways of governing.2 While criticizing transparency’s efficacy with one breath, with the next they defang it, claiming corporate secrecy exceptions will prevent useful information from getting out.3

The standard response to concerns about “black box” algorithms is to make those algorithms transparent or explainable. Such approaches, however, involve significant limitations, especially in professional contexts such as medicine, law, or financial advice. Instead, systems should be designed to be contestable, meaning that those subject to algorithmic decisions can engage with and challenge them. Both laws and norms can encourage contestability of automated decisions, but systems designers still must take explicit steps to promote effective questioning and challenges.