User interfaces are changing, but web search remains
It feels like every question under the sun can now be answered by an AI chatbot. Does that mean search engines, as a technology and as a way to explore the internet, are dying? The answer is no, but we are living through a significant transformation in information retrieval. Interfaces and search algorithms are evolving, but the fundamental need to seek and find information remains – along with the requirement for abundant, up-to-date data to enable discovery.
The Open Search Symposium (Ossym25) was held on October at CSC – IT Center for Science. The research and demos presented there show that AI not only supports the technical development of search engines but fundamentally changes how information is sought, interpreted, and utilized.
Ossym25 continued the annual conference series organized by the Open Search Foundation and its partners, aiming to promote independent European search technology. Many presentations were linked to the OpenWebSearch.eu project, which is developing an open European web search index – the Open Web Index (OWI) – and the supporting infrastructure.
Integrating AI into web search
One of the key technical trends identified at Ossym is the integration of generative AI into web search. Large Language Models (LLMs) can produce summaries, answers, and contextualized results directly from OWI data. LLMs not only improve the relevance of search results but also make information retrieval more understandable and efficient for users.
This approach is supported by Retrieval-Augmented Generation (RAG), where LLMs first retrieve relevant information from the index and then use it to support generated content. In this way, AI does not act merely as a text generator but combines search technology and generative capabilities into an intelligent whole.
European web index data (OWI) is available for AI development through the LUMI AI Factory’s Dataset-as-a-Service offering. One of the key goals of the LUMI AI Factory is to bring high-quality, valuable datasets into the realm of AI development and high-performance computing.
The LUMI AI Factory is the largest AI factory in the European Union. It serves as a joint platform and accelerator for RDI activities by companies, research groups, and public actors, enabling the development of innovative European AI models and applications. It seamlessly combines everything needed for AI development into one place: world-class computing power, high-quality data, and top-tier support and services. LUMI AI Factory’s services are free for startups, SMEs, and academic researchers.
Toward better search results
AI also plays a major role in enriching semantic information. Entities and their relationships can be automatically identified from web content, creating knowledge graphs. These enable search engines to understand content better and provide semantically relevant results to users.
Automatic detection of location data and topics is another practical application. AI can identify contexts such as places and themes in web content, enabling locally targeted and thematically refined search results.
AI also demonstrates its power by combining information from different sources to support innovation. At Ossym25, methods were presented for using LLMs to expand and merge technical data from sources such as patent registries and scientific research results, helping discover new connections. Keynote speaker Harri Ketamo (Headai) illustrated this potential by showing how a profile can be built from one’s CV and publicly available data, compared against job requirements, and then used to find targeted courses and training to match available positions.
Regulatory and ethical perspectives
The use of AI in web search is not just a technical issue – it also has economic, political, and societal dimensions, as well as implications for individual rights such as privacy.
European Commission Vice-President Henna Virkkunen sent a video greeting to participants, emphasizing the importance of open search for democracy. Member of the European Parliament Aura Salla, in her keynote, challenged the Commission’s AI policy priorities, stressing the need to create conditions for market-based solutions and reduce regulatory burdens to strengthen European companies’ global competitiveness.
Viivi Lähteenoja (MyData Company) highlighted the importance of epistemic virtue. Virtue ethics offers tools for assessing the reliability of information and identifying biases and misinformation. This ethical perspective is crucial when building future search ecosystems that serve society comprehensively.
The 2026 Ossym will take place in Berlin in October.
Author: Heidi Laine, WP leader for Data access and integration at the LUMI AI Factory at CSC.
Photo: Mikael Kanerva, CSC