The key question regarding responsible AI we deal with at the D4P is how to ensure that AI is not only legal and ethical by design, but also—crucially—liable and sustainable by design. Part of our answer is AI which is risk-aware and transparent: it recognises risks like bias and data manipulation, and calls for quality benchmarks, ethical guidelines, and reproducibility standards. In addition, we support focus on skills and data access, i.e. investment in upskilling researchers, improving access to data and computational resources, and supporting AI research communities.  

While responsible AI is of utmost importance, the environmental cost of AI is also becoming a major concern.  As AI becomes deeply embedded in digital infrastructures across all industries, the demand for scalable and sustainable systems is rapidly accelerating. Efficient AI directly addresses this need by minimising the energy footprint, computational costs, and latency across the entire AI lifecycle—from data processing and training to real-time inference—all without compromising performance or accuracy. This efficiency is vital not only for widespread adoption across sectors and society, but also for enabling reliable AI on edge devices, mobile platforms, and in resource-constrained environments.  D4P is committed to driving innovation in AI efficiency, creating AI technologies that are not only powerful but also sustainable.  Our core mission is to minimise AI’s environmental footprint through enhanced energy efficiency at every level.  By engineering AI systems that are both high-performing and resource-aware, we guarantee reliable operation on diverse devices while minimising their environmental impact.