Share This Article
The European Commission has recently published an updated version of the FAQs on the Data Act, offering further clarification on the concepts of inferred or derived data.
The Data Act, a key component of the European Data Strategy, is a Regulation designed to enhance the EU’s data economy and promote a competitive data market. It seeks to make data—particularly industrial data—more accessible and usable, fostering data-driven innovation and increasing data availability.
Inferred and derived data under the Data Act
The Data Act’s data-sharing obligations primarily apply to raw and pre-processed data. Raw data refers to data automatically generated without any additional processing. Pre-processed data, on the other hand, includes data collected from a single sensor or a network of sensors, processed to determine a specific physical quantity or quality—such as temperature, pressure, liquid level, position, acceleration, or speed—so that it becomes comprehensible for broader use cases.
However, inferred and derived data are explicitly excluded from the Data Act’s sharing obligations. This means that companies are not required to share such data with users or third parties. A key issue in this context is the distinction between inferred/derived data and pre-processed data.
The latest version of the FAQs provides further clarity on this matter, emphasizing that the level of data enrichment plays a crucial role in determining whether data falls under the sharing obligations. Under the Data Act, users have the right to receive, use, and port data they have (co-)generated. This right extends to raw and pre-processed data, including associated metadata. However, the Data Act also aims to protect incentives for investing in data technologies that enhance or transform data, provide additional insights, or enable autonomous decision-making.
To differentiate between raw/pre-processed data and inferred/derived data, the Data Act refers to key factors such as: “substantial modification“, “substantial investments in cleaning and transforming the data“, and “proprietary and complex algorithms“. Therefore, according to the updated FAQs, companies should assess whether their data has undergone substantial modifications or significant investments in data processing, particularly through proprietary algorithms, to determine whether it qualifies as inferred or derived data.
Moreover, according to the new version of the FAQs, the Data Act’s objective is to enable processing of data by a wide array of actors in the data economy and therefore shared data should be easily usable and understandable by entities other than those who generated it. At the same time, the need to interpret raw data should not translate in an obligation on the data holder to make substantial investments in cleaning, transforming or reformatting data to make it understandable. Instead, the European Commission expects users and third parties to have a reasonable level of technical competence to interpret the data provided.
The data sharing requirements under the Data Act
With the Data Act set to be applicable (i.e., on 12 September 2025), companies providing connected products and related services must assess their compliance with its data-sharing requirements. This includes evaluating whether data obtained, generated, or collected through their products or services qualifies as raw, pre-processed, or inferred/derived data.
Furthermore, the Data Act includes additional exemptions, such as cases involving trade secrets or specific security requirements. Companies must conduct a thorough analysis of the categories of data they process to determine their obligations under the Data Act and, where necessary, they should establish appropriate mechanisms for granting either direct or indirect access to data in compliance with the regulation.
On the same topic, you can read the following article “Data Act approved: What changes on IoT trade secrets protection“.
