Experts have warned that “silence” on new data guidelines risks hampering fair competition in AI development and allowing dominant companies to ramp up artificial intelligence.
Researchers believe the rules create uncertainty about the use of data and risk giving some companies preferential access.
In response to the consultation on the Joint Guidelines on the Interaction of the Digital Markets Act and the General Data Protection Regulation, the Competition Shaping Competition in the Digital Age (SCiDA) team said the lack of a policy on AI training creates enforcement uncertainty and risks enabling practices that may clearly violate the DMA’s restrictions on data combination and interuse.
The SCiDA team says ignoring AI services in shaping digital markets is a fundamental mistake, as it means a key driver of competition will fall outside the scope of regulatory oversight.
Without clear requirements for AI training and deployment, gatekeepers are free to combine large datasets to train superior models, but competitors do not have access to comparable data, creating structural barriers to effective competition in AI development.
Responses to the consultation were provided by Oles Andreychuk, Pavlina Hubkova, and Anush Ganesh from the University of Exeter, Ruprecht Pozsun Kena Jen and Sarah Hink from the Heinrich-Heine University of Düsseldorf, and Jasper van den Boom from Leiden University.
The study states that while the guidelines represent an important step towards consistent GDPR-DMA application, there are significant gaps that require clarification.
Conflicts arise whenever the GDPR allows for an interpretation that allows DMA circumvention.
Professor Pozsan said: “As the Competition and Markets Authority has warned, the lack of clear guidelines could create a feedback loop in which companies are able to strengthen their position within the digital realm, generating rich data sets with favorable access to AI, which in turn could lead to greater access to the data needed to build and improve underlying models.”
Professor Andreichuk said: “The guidelines are not sufficient to define standards regarding anonymity. This means that companies that are “gatekeepers” of data have a “blank slate” and companies that want greater access for competitive purposes do not. ”
The SCiDA team argues that gatekeeper companies adhere to prohibitively high data protection standards when doing so allows them to circumvent other laws, but warns that these standards can be completely ignored when it suits their business interests.
The purpose of the Guidelines is to provide guidance on some provisions of the DMA that relate to or may involve the processing of personal data by gatekeepers, or that include reference to GDPR concepts and definitions.
Some companies have invoked the data protection and privacy requirements protected by the GDPR as a justification for avoiding compliance with data-related obligations under the Digital Markets Act. The consultation response states that the GDPR should not be compromised in pursuit of the objectives of the DMA, but that the objectives of the DMA should not be undermined by the application of other legal frameworks such as the GDPR or the ePrivacy Directive.
Dr. Habkova said: “If designations under the DMA persist in a narrow, platform-centric model of the digital economy, the regime risks becoming obsolete. A more effective approach would incorporate a broader theory of harm that addresses the evolving structures of power and dependence in AI-driven markets.”
The SCiDA team recommends new guidance that where the GDPR allows for multiple interpretations, the “gatekeeper” should choose the one that least impedes the objectives of the DMA and maintains GDPR compliance.
/Open to the public. This material from the original organization/author may be of a contemporary nature and has been edited for clarity, style, and length. Mirage.News does not take any institutional position or position, and all views, positions, and conclusions expressed herein are those of the authors alone. Read the full text here.
