DEVELOPMENT OF AI GOVERNANCE MODEL IN ENTERPRISES BASED ON CMMI MODEL STRUCTURE (STUDY CASE: HEALTHCARE DOMAIN)

• Main Author: Luthfi Ramdhani, Yusuf
• Format: Theses
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/79483
• Institution: Institut Teknologi Bandung

The adoption of Artificial Intelligence (AI) in enterprises often poses new challenges in the enterprise and society. To deal with these challenges, we need an AI governance that can control and direct AI in the enterprise well. However, there is no existing AI governance model that has a number of measurable practices to properly control and direct AI in enterprises. Capability Maturity Model Integration (CMMI) can be used as a reference to develop an AI governance model with a structure that has a number of measurable practices. As for the content of this AI governance model, a number of examples of AI governance practices in the healthcare domain will be identified. Thus, this study will discuss the development of AI governance model in enterprises based on the CMMI model structure, with a focus on model content in the healthcare domain. This AI governance model from now on will be called Model Tata Kelola AI Terukur (MTAIT). The MTAIT model design phase produces a model structure consisting of layer, capability area, practice area, practice level, and practice, and produces the basic content of the model consisting of data layer, technology layer, ethics & social layer, and legal layer. The MTAIT model development phase produces model content that is fully broken down from the layer level down to the practice level, as well as a way of measuring the practices by determining the capability level at the practice area level and maturity level at the capability area, layer, and up to the enterprise level. Based on the evaluation results on the MTAIT model, it can be concluded that the MTAIT model is in accordance with the specified AI governance model specifications, can be used to a practical level and be measured, and can help to control and direct AI in the enterprise, but not yet fully in accordance with the needs of AI governance practices in healthcare.