Drivers of digital technologies

Last updated: 14.8.2026

Image

Computer-implemented inventions

Computer-implemented inventions (CII) are technological solutions in which at least one feature is executed by a computer program running on a computer, network or programmable device.

CII are a key enabler of modern technology. They:

• integrate software algorithms with hardware functionality
• enable advancements in automation, data processing and computational efficiency
• support real-time data processing and the optimisation of complex workflows

Through embedded software, sensor technologies and advanced computing, CII support:

• automated processes and adaptive decision-making
• digital transformation across manufacturing, healthcare, transportation and enterprise solutions

Artificial intelligence (AI) relies on software-driven decision-making models, while quantum computing (QC) utilises complex algorithms and simulations. Accordingly, both AI and QC are based on CII, which provide the essential computational frameworks that enable these solutions to function effectively.

CII are driving innovation across multiple industries. For instance:

• pharmaceutical manufacturing: 
  Precision dosing and automated quality control systems use sensor-driven measurement technology to optimise production accuracy (e.g. EP3487685B1)
• autonomous vehicles: 
  AI-powered sensor fusion and machine learning algorithms enable real-time decision-making and safe navigation (e.g. EP3707572B1)
• Industry 4.0
  The ongoing transition from digital automation to a connected, intelligent manufacturing environment, uses internet of things (IoT) connectivity, real-time analytics and AI-driven automation to improve production efficiency and enable predictive maintenance (e.g. EP3717918B1)
• Digital infrastructure
  Computer-aided design (CAD) coupled with simulations and digital modelling allow industries to create product prototypes, optimise engineering processes and refine complex structures before physical production (e.g. EP2013798B1)

Advances in video compression

​​Polish software engineer Marta Karczewicz dedicated her career to inventions that enable fast, high-quality streaming. Named as inventor in almost 130 granted European patents, Karczewicz's inventions have transformed the video entertainment industry and made video streaming on laptops and mobile devices available to a wider audience. Marta Karczewicz was a European Inventor Award 2019 finalist in the Lifetime Achievement category.

Artificial Intelligence

Artificial intelligence (AI) systems are built on machine learning (ML) techniques, which enable algorithms to:

• learn from data
• recognise patterns
• optimise decisions
• autonomously refine their performance

AI learns from data in different ways.

• Supervised learning involves training an AI model on labelled datasets in which the correct outputs are known, enabling it to make accurate predictions. 
• Unsupervised learning works with unlabelled data, meaning AI groups similar information together or finds hidden trends and patterns without prior knowledge. 
• Reinforcement learning allows AI to learn by interacting with its environment, improving its decisions over time based on past experiences. Instead of relying on predefined rules, AI adapts and refines its approach through trial and error.

Deep learning is a branch of machine learning that mimics the way the human brain processes information. It uses artificial neural networks, consisting of layers of interconnected "neurons" that process information in stages. These networks enable AI systems to perform complex tasks such as understanding spoken language, recognising images and analysing sensor data. To improve accuracy and refine its predictions, the system continuously adjusts its internal settings using specialised learning techniques.

Use of AI across industries

AI has evolved from theoretical research into an integral part of multiple industries, including healthcare, finance, automation and cybersecurity. In everyday life, for instance, AI enhances smartphone camera image processing by adjusting lighting and sharpening details to improve photo quality. AI also powers voice assistants, enabling seamless interaction through speech recognition and real-time language comprehension. In medical diagnostics, machine learning algorithms analyse MRI scans to detect abnormalities more quickly and accurately, helping doctors identify diseases at an early stage (e.g. EP3659494B1, relating to AI-assisted medical imaging analysis for disease detection).

Milking robots for healthier cows

An automated system designed by Dutch duo Alexander van der Lely and Karel van den Berg allows cows to decide when it's time to be milked. The invention keeps animals happy and healthy, increases milk production and equips farmers with diagnostic tools to operate more efficiently. Alexander van der Lely and Karel van den Berg were European Inventor Award 2019 finalists in the Industry category.

AI brings earlier care to new Parkinson's patients

After watching a YouTube clip of Parkinson's patient Michael J. Fox, US-based Erin Smith developed an AI-powered app that uses video to enable early detection of Parkinson's disease. Her invention can help to slow development of the condition. Erin Smith was a winner of the Young Inventors Prize 2022.

​​A brain implant for two-way communication with external devices​

​​Thomas Oxley and Nicholas Opie invented the Stentrode, an implantable brain-computer interface (BCI) that enables patients living with paralysis to communicate through thought. Thomas Oxley and Nicholas Opie were selected as finalists for the European Inventor Award 2023.

Generative AI

One of the most significant developments in this area is generative AI, which enables machines to generate entirely new content, including text, images, music and software code, going beyond the analysis of existing data to produce novel outputs. This is achieved through deep generative models, such as generative adversarial networks (GANs), which learn by pitting a generator against a discriminator, and variational autoencoders (VAEs), which learn compact representations of data to produce new samples. Autoregressive models, such as Generative Pre-trained Transformer (GPT), take a different approach by generating content step by step, predicting each element based on prior context. These models are particularly effective in tasks such as text generation, where each word or sentence is created sequentially to maintain coherence and logical flow.

Significant advancements in generative AI have been driven by transformer-based architectures such as GPT, Large Language Model Meta AI (LLaMA) and DeepSeek (e.g. EP4152314B1, relating to context-aware AI text generation). These architectures use self-attention mechanisms to efficiently process complex relationships in data, enabling AI to produce structured, coherent, context-aware outputs, making them invaluable across a variety of fields. Other transformer-based models, such as Bidirectional Encoder Representations from Transformers (BERT), focus on language understanding rather than generation, excelling at tasks like text classification, search and information extraction. In software development, for instance, generative AI can assist with automated code generation and debugging, thereby reducing programming time. In the pharmaceutical industry, AI-driven simulations can accelerate drug discovery by predicting molecular interactions. The ability to generate high-quality synthetic media supports innovation in design and content creation by automating creative processes accurately.

Autonomous AI systems

Autonomous AI systems are transforming mobility and engineering. Self-driving vehicles use computer vision, sensor fusion and deep learning algorithms to navigate dynamically, recognise obstacles and optimise driving routes (e.g. EP3947095B1, relating to real-time decision-making in autonomous driving). AI-powered predictive modelling enhances safety by analysing road conditions and adjusting vehicle behaviour accordingly. In industrial automation, AI-driven autonomous design systems generate optimised mechanical and structural models, reducing material waste and enhancing functionality. These developments are setting new standards in precision engineering, robotics and automated manufacturing.

Quantum computing

Quantum computing (QC) redefines computing system capabilities. Unlike classical computers, which use a binary system of ones and zeros to switch transistors on and off, quantum computers use qubits. Based on principles of quantum mechanics, such as superposition and entanglement, qubits process information in entirely new ways. Quantum algorithms are expected to deliver drastic speed-up advantages (polynomial to exponential) over classical algorithms commonly used for similar problems or tasks. As more qubits become entangled, the computing power of quantum computers increases exponentially, making them a key component of next-generation technology.

Quantum computers are not designed to replace classical computers, but rather to complement them. By working together, both technologies can address problems that conventional systems struggle with.

Despite its vast potential, quantum computing faces significant technical challenges. Qubits are highly sensitive and prone to disruption by external factors such as temperature fluctuations, stray electromagnetic fields or cosmic radiation. These disturbances can cause qubits to lose their quantum state, forcing them into a conventional binary form. Researchers and engineers are continuously developing strategies to stabilise qubits and enhance system reliability.

Quantum computing is already being explored and implemented in various domains, including both hardware-related innovations and software-based approaches:

• quantum computing system with improved error mitigation (focused on physical qubit architecture and stability, e.g. EP3602423B1) – This open-source software allows programmers and researchers to build quantum algorithms and applications.

• quantum annealing system for solving complex computational tasks (hardware-based quantum optimisation, e.g. EP3754566B1) – This solution is used for protein folding simulations, portfolio management and logistics optimisation.

• quantum key distribution system for secure data transmission (quantum cryptography software and algorithms, e.g. EP2697931B1) – This solution focuses on efficient key management and distribution to provide robust encryption and protect against evolving cyber threats. It is ideal for sectors requiring highly secure communications, including financial services, government agencies and healthcare organisations handling sensitive data.