How digital platforms respond to regulation

Regulatory changes may pose headwinds to platform business models. How do we address such headwinds?

This article is part of the Digital Platforms hub

Firms today execute platform strategy in an ever- changing regulatory environment. On one hand, centuries-old regulation written for industrial-era competition is increasingly being rewritten. On the other hand, many platform firms have operated in a regulatory
Wild West over the course of the first two decades of the twentieth century. As regulators catch up, these firms have to revisit their business model assumptions to retain their competitive advantage in the face of regulatory headwinds.

Consider, for example, the increasing regulation around data privacy. Between 2005 and 2015, most platforms harvested and exploited user data, operating in a regulatory lacuna. As privacy regulations like the GDPR (General Data Protection Regulation) in the European Union gain greater ground, companies whose business models relied heavily on unrestricted access to user data need to revisit and reshape their business models.

In general, the GDPR places a growing compliance burden on firms which capture and leverage user data. These businesses, in turn, need to restructure their machine learning and AI efforts to ensure their platform efforts comply with changing regulation.

Feel Free to Share

Since the rise of the GDPR, firms have structured different approaches to comply with new regulation while retaining their ability to leverage user data for competitive advantage.

To comply with the GDPR, some firms use federated learning. With federated learning, personal data doesn’t leave the system of residency/storage. Instead, an AI model trains locally on each individual system with local data. This trained model is then merged with the master AI model on a different system. This ensures GDPR compliance since data doesn’t leave the system of storage. This does come with certain limitations as the locally trained AI model is limited in its ability to inform the master model, compared with using the entire data input to train the master model.

Some platform firms also use transfer learning, which utilizes an existing learning model and retrains itself using that model towards a new use case. With a well-trained library of models, firms can reduce their dependence on new data capture, when expanding to new use cases.

To further reduce dependency on capture of usage data, some platform firms leverage Generative Adversarial Networks (GAN) to generate input data with the help of output data. Two neural networks are used in parallel to train the model – the first acts as a generator that predicts the output and the second acts as a discriminator that determines the difference between real output data and generated output data. GANs can be very powerful when used in combination with federated and transfer learning.

As these examples above demonstrate, platform firms need to invest heavily in complying with a changing regulatory landscape, to ensure that they can compete effectively while accounting for regulatory headwinds

👈 Previous article related Digital Platforms            Next article related to Digital Platforms👉

Frequently Asked Questions

How can platform firms adapt to changing regulations like GDPR while using user data to maintain competitiveness?

Platform firms face the challenge of adapting to regulatory changes such as the GDPR, which imposes stringent requirements on data privacy and usage. To address these challenges, firms are restructuring their business models and adopting innovative approaches to ensure compliance while still leveraging user data for competitive advantage. For example, federated learning enables AI models to be trained locally on individual systems without transferring personal data, thus ensuring GDPR compliance. Transfer learning allows firms to retrain existing models for new use cases, reducing the need for additional data capture. Additionally, Generative Adversarial Networks (GANs) help generate input data from output data, further reducing dependency on capturing user data while complying with regulations.

What are the specific challenges that platform firms face in complying with regulations like the GDPR, especially considering their reliance on user data for competitive advantage?

While these approaches offer solutions to GDPR compliance, they also come with challenges and limitations. Federated learning, for instance, restricts data access to comply with GDPR regulations, which may limit the effectiveness of locally trained AI models compared to using the entire dataset. Transfer learning, although reducing the need for new data capture, may encounter difficulties in adapting to diverse use cases or domains, potentially impacting model performance. Thus, platform firms must carefully assess the trade-offs and limitations associated with each approach to ensure effective regulatory compliance without compromising their competitive edge.

How do platform firms balance regulatory compliance and effective use of user data, employing strategies like federated learning and GANs?

In addition to federated learning, transfer learning, and GANs, platform firms are exploring a range of strategies to navigate the regulatory landscape while maximizing the value of user data. These strategies may include implementing techniques such as differential privacy to anonymize data, enhancing encryption methods to protect sensitive information, and developing robust data governance frameworks to ensure compliance with regulations like GDPR. By employing a combination of these strategies, platform firms aim to strike a balance between regulatory compliance and maintaining their ability to harness data for innovation and competitive advantage in the digital ecosystem.