The AI Assistants for Knowledge Engineering (AIAA4KE) 2026 workshop explores the potential of generative AI and agentic workflows in Knowledge Engineering. While AI-assisted approaches have revolutionised general software engineering, and ontologies and knowledge graphs (KGs) have been recognised as valuable components of AI systems, knowledge engineering -- including ontology modelling and knowledge graph construction -- remains a bottleneck, reliant on scarce human expertise.

The workshop covers the use of generative AI for knowledge engineering and AI assistants, as well as related agentic systems, with a focus on industrial applications, use cases, and data. Submissions regarding the following topics, but not limited to, are especially welcomed.

AI assistants for ontology modeling & matching, involving:

• User intent analysis based on unimodal or multi-modal inputs
• Different interaction paradigms (single-shot, chat-based, or, wizard)
• Ontology and taxonomy generation
• Ontology matching
• LLM-based ontology reasoning and reasoner-in-the-loop approaches.

AI assistants for knowledge graph construction, curation, and validation:

• AI assistants for KG generation and population from unstructured, semi-, and structured data sources
• AI assistants for constructing KGs from modalities beyond text
• AI assistants for generating and processing SHACL constraints
• AI assistants for generating and processing SPARQL queries with focus on transformations.

Evaluation, benchmarks and governance of AI assistants:

• Emerging standards (e.g., MCP or A2A) in the realm of AI assistants and their relation to the Semantic Web stack
• Shared tasks and benchmark and data sets for modeling assistants, including ideas on joint publications of ontologies and data sets across companies
• Benchmarks and datasets for AI model based understanding of ontologies/KGs
• Novel benchmarking methodologies such as human based evaluation, LLM as a judge, dynamic benchmarks
• Evaluation metrics/approaches for the evaluation of usefulness of ontologies/KGs generated via AI models
• Methods for logging actions of AI assistants, as well as audit and rollback in knowledge engineering tasks
• Patterns for keeping human-in-the-loop on high impact changes, including approval and governance workflows.

Adoption of AI-based knowledge engineering:

• Industry and in-use use cases for AI assistants that support knowledge engineering tasks
• Scalability and resource-efficiency of AI-based knowledge engineering
• Open challenges in real-world application scenarios

Submission Guidelines

We welcome contributions on the above topics that cover the RDF/OWL technology stack, labeled property graphs, or a combination of both and invite the submission from the following categories:

• Full papers (max. 12 pages)
• Short papers (max. 6 pages)

Page limits apply excluding the references.

Lightning talks: Submitting a talk for position, vision, and challenge statement is highly encouraged. These talks will be 5min, required a one-pager (submitted via Easychair) to describe the topic. The topic can/will be picked up for a plannend panel discussion.

Submissions have to be submitted electronically using EasyChair.

Committees

Organizing Committee

• Maribel Acosta, TU München, Germany

• Heiko Paulheim, Universität Mannheim, Germany

• Evgeny Kharlamov, Bosch GmbH, Germany & University of Oslo, Norway

• Diego Rincon-Yanez, Vienna University of Economics and Business, Austria

• Felix Sasaki, SAP SE, Germany

• Patrik Schneider, Siemens AG, Germany & TU Wien, Austria

Contact

All questions about submissions should be emailed to: patrick-schneider@siemens.com