A new peer-reviewed study has found that while artificial intelligence (AI) and digital tools could deliver significant energy savings in European buildings, regulatory uncertainty is delaying adoption.
Published in Buildings (June 2025) and based on 64 sources, the research examines how EU regulations influence the use of AI and Internet of Things (IoT) technologies in building energy systems. It highlights how privacy rules, cybersecurity requirements, and liability frameworks are shaping investment decisions.
The study reviews legislation including the AI Act, the Energy Performance of Buildings Directive (EPBD), and the General Data Protection Regulation (GDPR). It identifies a gap between EU ambitions for smarter, more efficient buildings and the limited deployment of available digital solutions.
Researchers point to three main challenges: legal uncertainty over compliance, technical barriers such as interoperability and explainability, and the continued undervaluation of the economic benefits of digital retrofits. According to the study, AI-based tools can cut energy consumption by 20–40%, but uptake remains limited.
Industry experts suggest that AI could become one of the most effective ways for building owners to meet EU climate targets. Donatas Karčiauskas, CEO of energy optimisation company Exergio, said that many owners are hesitant due to compliance concerns but stressed that solutions exist to ensure both efficiency and regulatory alignment. “With the right architecture – such as edge computing, anonymised data, and explainable AI – it is possible to meet requirements and still achieve results,” he said, noting that some projects recoup costs in under a year.
The study also highlights the role of advanced methods such as federated learning, which allows AI models to train locally without sharing sensitive data, and SHAP tools that explain algorithm outcomes. These, the authors argue, address many concerns around trust and transparency.
Digital retrofits, which use AI to adjust building systems in real time based on occupancy, weather, and energy prices, were found to deliver greater flexibility than traditional upgrades. However, current policies often treat these systems in the same way as basic automation, limiting their wider use.
While regulation can act as a barrier, the research concludes that it also acts as a driver by raising standards for security and system compatibility. Framing AI as both a performance enhancer and a compliance pathway could accelerate its adoption, the authors argue.
With key EU policy deadlines approaching in 2025, the report suggests that early adoption of AI-enabled energy systems will be critical to meeting decarbonisation goals.
Read the full report HERE.
