Algorithmic transparency in rankings - a possible expansion to personalisation?
In the digital economy, algorithms are the new air- they are everywhere, essential for survival online. Algorithms manage every keystroke, every search, every step on the web. Algorithms are a sequence of instructions to solve a problem and take decisions. Online platforms use algorithms in every step of their operations, from search results to payment portals.
Though executed by machines, these algorithms are created by humans and awarded intellectual property protection. In modern online transactions, algorithms form an integral part of any online platform’s business model and their competitive advantage in the market.
One of the ways that algorithms impact consumer choice and affect decision-making is by ranking products and services on the platform. Consumers tend to choose the products first listed or ranked higher over others. As platforms became aware of this consumer behaviour, they began manipulating the appearance of listing by either favouring their products or selling the top listings to the highest bidder, thereby affecting competition in the market.
In the quest for protecting consumers, legislations call for transparent information for consumers who make rational decisions based on their information. One of the ways to achieve this transparency is- algorithmic transparency.
Algorithmic Transparency refers to the principle of disclosure of factors that influence algorithms to those who are affected by the systems that employ these algorithms
As consumers are impacted by, and the algorithms use their data, transparency was called for.
In the EU, in an attempt to achieve this balance, the provision of information and the protection of intellectual property of businesses in rankings, the Fairness and Transparency Regulation(P2B), introduced algorithmic transparency between platforms and business users. This Regulation was supplemented by the Ranking Guidelines, which served as a guide to platforms to disclose the main parameters that the algorithm considers to provide these rankings. The Regulation and the Guidelines attempt to create a level playing field amongst all the competitors in the market. The Modernisation Directive also included this provision, which amended the Unfair Commercial Practices Directive. The Proposed Digital Services Act has also included algorithmic transparency about the main parameters used in creating recommendation systems.
Similarly, in India, the Consumer Protection (E-Commerce) Rules state that platforms must be transparent about the main parameters used by the algorithm in ranking products and services. The Indian Consumer Protection (E-Commerce) Rules are comprehensive and cover all marketplaces, being very technology-neutral. On the other hand, the EU’s Fairness and Transparency Regulation applies only between platforms and business users. Consumers are covered under the broader ambit of the Modernisation directive and the UCPD
Legislation in the EU and India are focused on algorithmic transparency concerning rankings. Algorithmic transparency seeks to balance the protection of users’ rights and businesses’ intellectual property rights. With the provision of transparency information about algorithms, consumers can make rational decisions without the disclosures of the algorithm, which is an integral part of a platform’s business model.
Algorithmic transparency provides information to consumers about the data that is input into the algorithm without revealing the algorithm. Algorithmic transparency attempts to balance the protection of businesses’ intellectual property and the needs of consumers. As algorithmic transparency walks this tightrope, it raises ideas as to whether a similar standard can be applied to other aspects of platform businesses that operate based on algorithms. For example, a similar algorithmic transparency standard can be called for in providing information to consumers about the parameters that create personalised prices online.
Personalised prices are created to cater to consumers’ needs based on collected consumer data. Personalised prices are created using data processed by algorithms. Drawing parallels with the ranking transparency model, the main parameters could also be provided to consumers, where prices are personalised. This algorithmic transparency in personalisation would provide consumers with the information about the collection and use of their data to provide a personalised price and help them make an informed decision.
Algorithmic transparency is one of the ways to pierce the algorithmic veil empowering corporations to collect, store and use consumer data while balancing business and consumer needs and rights.
|This guest blog was written by Pratiksha Ashok for the IGIR and METRO Faculty of Law Maastricht #COMIPinDigiMarkts2022 project - More blogs on Law Blogs Maastricht|
This guest blog is part of the project #COMIPinDigiMarkts2022. These blogs have been specially prepared by participating internal and external project members and focus on competition law and IP law, with particular reference to the digital markets.