M4ESTRO’s portfolio of Manufacturing as a Service solutions now gears up with a new addition, the Human-Centric Decision Support System (HC-DSS). This tool gives visibility into why an AI system makes certain recommendations. This explainability is key as organizations scale their AI deployments and seek to build internal trust in these systems.
When receiving recommendations from AI systems, users may not trust the results and often feel the need to double-check them. This problem is intensified by the rapid expansion of AI and its increasing integration into traditional processes and methods. HC-DSS is a new tool that transforms recommendations from AI models into transparent and practical outcomes for manufacturing professionals, giving them visibility into why a model made certain suggestions. This helps demystify AI decision-making and build trust in these systems.
To achieve this, the HC-DSS combines several explainable AI techniques (LIME, counterfactual explanations) with human expertise to provide outcomes that help users better interpret AI results and explore suitable alternatives based on reasoning and recommendations. The HC-DSS supports human oversight, minimizing potential risks while strengthens human understanding.
In one of M4ESTRO’s use-cases, the FAE-IMECH use case, HC-DSS was integrated with other developments within M4ESTRO, such as the Optimisation Engine, a tool that assesses the most efficient and resilient value chain paths, and the AI assistant, a NLP chatbot. This integration demonstrated how AI results can be interpreted while the combination with natural-language interaction provided by the integration of the NLP module offers an additional level of explainability. The next step is to bring HC-DSS closer to real manufacturing environments through integration, testing, and validation in all M4ESTRO use-cases.
The development of a Human-Centric DSS is an important addition to M4ESTRO and the manufacturing sector in general, where the use of AI systems and automation is continuously increasing. The HC-DSS helps users better interpret recommendations and helps them make more informed decisions which are better thought-out, while minimizing associated risks. These are vital ingredients for a successful and wide AI deployment built on trust.
