Ubie, as JaDHA Working Leader, Collaborates with AI Safety Institute (AISI) to Formulate the 'AI Safety Evaluation Guideline in the Healthcare Domain'
The Japan Digital Health Alliance (JaDHA), with Ubie as the working leader, has established an AI safety guideline for the healthcare sector in collaboration with AISI to safely implement generative AI technologies.
📋 Article Processing Timeline
- 📰 Published: April 3, 2026 at 20:06
- 🔍 Collected: April 3, 2026 at 11:30
- 🤖 AI Analyzed: April 21, 2026 at 03:53 (424h 22m after Collected)
The Japan Digital Health Alliance (hereinafter 'JaDHA'), where Ubie, Inc. (Headquarters: Chuo-ku, Tokyo, Co-CEOs: Yoshinori Abe, Kota Kubo, hereinafter 'Ubie')—whose mission is 'to guide people to appropriate medical care with technology'—participates as a working leader company, has formulated the 'AI Safety Evaluation Guideline in the Healthcare Domain' (hereinafter 'this Guideline') in collaboration with the AI Safety Institute (hereinafter 'AISI'). This aims to accelerate the safe social implementation of generative AI technologies, including Large Language Models (LLM), in the healthcare field.
This Guideline provides specific evaluation methods for practitioners, reflecting the sensitivities and risks unique to medical care and healthcare, referencing the evaluation perspectives formulated by AISI.
AISI: Press Release
## Background and Objectives
In recent years, generative AI has brought a 'once-in-a-decade innovation' to the healthcare sector, improving operational efficiency for doctors and supporting patient communication. On the other hand, there are mounting challenges unique to the healthcare domain, such as the risk of health hazards due to hallucinations (generation of misinformation), the need for advanced privacy protection, and ensuring security.
As discussed at the 'Hiroshima Global Forum for Trustworthy AI' held in January 2026, building 'Trustworthy AI' has become the most critical issue in the international community.
Ubie led the formulation of this Guideline as the working leader of JaDHA's Healthcare SWG. We have aimed to create an environment where businesses can ensure safety from the development and design stages, achieving both business value and peace of mind/safety.
## Main Features of the Guideline
This Guideline has the following features so that even companies with few experts can easily utilize it.
1. Setting up evaluation methods at five development and design stages aligned with the AI lifecycle
- Product Design: Clarification of product objectives/use cases, risk assessment, construction of governance structure
- Model Selection: Selection of models suitable for the intended use and safety evaluation
- Product Implementation: System architecture, prompt engineering, guardrail implementation
- Product Verification: Comprehensive testing/verification and risk assessment
- Product Introduction and Operation: Monitoring and continuous improvement in the production environment
2. Multifaceted evaluation perspectives of 10 items and their respective specific risk assumptions
- Output control of harmful information: Related to medical and health...
This Guideline provides specific evaluation methods for practitioners, reflecting the sensitivities and risks unique to medical care and healthcare, referencing the evaluation perspectives formulated by AISI.
AISI: Press Release
## Background and Objectives
In recent years, generative AI has brought a 'once-in-a-decade innovation' to the healthcare sector, improving operational efficiency for doctors and supporting patient communication. On the other hand, there are mounting challenges unique to the healthcare domain, such as the risk of health hazards due to hallucinations (generation of misinformation), the need for advanced privacy protection, and ensuring security.
As discussed at the 'Hiroshima Global Forum for Trustworthy AI' held in January 2026, building 'Trustworthy AI' has become the most critical issue in the international community.
Ubie led the formulation of this Guideline as the working leader of JaDHA's Healthcare SWG. We have aimed to create an environment where businesses can ensure safety from the development and design stages, achieving both business value and peace of mind/safety.
## Main Features of the Guideline
This Guideline has the following features so that even companies with few experts can easily utilize it.
1. Setting up evaluation methods at five development and design stages aligned with the AI lifecycle
- Product Design: Clarification of product objectives/use cases, risk assessment, construction of governance structure
- Model Selection: Selection of models suitable for the intended use and safety evaluation
- Product Implementation: System architecture, prompt engineering, guardrail implementation
- Product Verification: Comprehensive testing/verification and risk assessment
- Product Introduction and Operation: Monitoring and continuous improvement in the production environment
2. Multifaceted evaluation perspectives of 10 items and their respective specific risk assumptions
- Output control of harmful information: Related to medical and health...