Joint Research Begins on AI-Powered 'Personalized Mold Forecast System' to Predict Indoor Mold Risks
Keio University SFC Research Institute and Hearts Rich Co., Ltd. have launched a joint research project starting in April 2026. The project aims to develop a smartphone and AI-based system to proactively predict and prevent mold generation risks.
📋 Article Processing Timeline
- 📰 Published: May 20, 2026 at 19:00
- 🔍 Collected: May 20, 2026 at 10:31
- 🤖 AI Analyzed: May 22, 2026 at 09:25 (46h 53m after Collected)
The Nakazawa-Okoshi Laboratory (Professor Jin Nakazawa) at the Faculty of Environment and Information Studies, Keio University SFC Research Institute, and Hearts Rich Co., Ltd. (CEO: Hideki Hokari) have launched a joint research project for the social implementation of an AI-powered "Personalized Mold Forecast System." This system aims to proactively predict mold generation risks in living spaces and facility environments—such as homes, elderly care facilities, hospitals, and warehouses—supporting preventive actions by residents, facility managers, and business operators.
Mold can occur not only in homes but also in various indoor spaces where people live, stay, and store goods, including elderly care facilities, hospitals, warehouses, and commercial facilities. Traditional reactive approaches, which involve removal and cleaning after mold has already developed, can lead to odors, deterioration of building materials and stored goods, anxiety among users, and increased facility management costs.
In this joint research, the partners aim to build an "Expert Knowledge Fusion AI" that estimates individual mold generation risks. This AI will integrate simple smartphone diagnostics, information on lifestyle habits and facility usage, location data, weather APIs, and over 10,000 cases of on-site knowledge related to consultations, surveys, and construction that Hearts Rich Co., Ltd. has accumulated to date.
Through this initiative, the goal is to shift mold countermeasures from "reacting after occurrence" to "predicting and preventing before occurrence," thereby contributing to improved hygiene management, storage environment control, and better living conditions in residential spaces and facility environments.
The research theme for this joint project is the "Social Implementation of Indoor Mold Risk Prediction and Behavior Modification Systems through the Fusion of Participatory Sensing and Expert Knowledge." The research period is scheduled to run from April 1, 2026, to March 31, 2028.
[Shifting Mold Countermeasures in Living Spaces and Facilities from "Post-Occurrence Reaction" to "Predictive and Preventive Management"]
Mold occurring in indoor spaces is not just an issue of visible dirt and odor; it significantly impacts the quality of life and business operations by causing building material deterioration, quality degradation of stored goods, anxiety about the indoor environment, and burdens on facility management.
However, in ordinary households and facility management settings, it is difficult to determine "which locations are prone to mold," "how much risk the current indoor environment holds," and "whether action should be taken immediately." In many cases, countermeasures are only implemented after visible mold or odors have appeared.
Furthermore, while indoor environment measurement utilizing IoT sensors is an effective method, it faces challenges such as introduction costs, installation efforts, and hurdles to continuous use, making broad adoption in homes and small-to-medium facilities difficult.
Therefore, this joint research will explore a mechanism for predicting mold risk that is more accessible to a broader range of residents and facility managers by combining simple diagnostic information via smartphones, weather data, and the decision logic of experts.
[AI-Powered "Personalized Mold Forecast System" Fusing Smartphone Diagnostics, Weather APIs, and Expert Knowledge]
The "Personalized Mold Forecast System" targeted in this research is an AI-driven prediction tool that estimates individual indoor mold generation risks. It combines indoor environment data, lifestyle habits, and facility usage inputted by users on their smartphones with regional weather data and expert knowledge.
Instead of merely sending alerts like "It is mold season," the system aims to present specific recommendations on which locations require attention and what preventive actions should be taken, based on the space's purpose and usage conditions.
Consequently, residents and facility managers will be empowered to take appropriate actions—such as ventilation, dehumidification, cleaning, improving storage methods, and adjusting inspection frequencies—at the right timing within their daily management routines, thereby proactively suppressing mold generation and recurrence.
[Reflecting Field Knowledge of Mold Countermeasures into the System as an Expert Decision Tree]
Hearts Rich Co., Ltd. is a company that has been working on mold surveys, mold countermeasures, and indoor environment improvements for homes and facilities. Based on the knowledge accumulated through its past on-site responses, the company will organize the conditions prone to mold, location-specific risks, indoor environments where mold recurs easily, and their relationships with usage conditions and management methods. This information will be integrated into the system's decision logic.
In this research, this specialized knowledge will be structured as an "Expert Decision Tree." By combining it with smartphone diagnostics and weather data, the project aims to realize highly practical mold risk assessments grounded in real-world field experience.
[Considering Application to a Wide Range of Indoor Spaces Including Homes, Facilities, and Warehouses]
While this research primarily targets the Japanese living environment, it envisions future application to a broad spectrum of indoor spaces, including elderly care facilities, hospitals, warehouses, and commercial facilities.
The risk of mold generation is influenced by temperature, humidity, building structure, number of occupants, ventilation environments, types of stored goods, and cleaning frequency; thus, the points requiring attention vary depending on the space's purpose. Moving forward, while advancing verifications within Japan, the project will also examine the applicability of mold prediction models factoring in usage purposes and regional differences.
Additionally, the project is considering applications to high-temperature, high-humidity regions such as Southeast Asia.
Mold can occur not only in homes but also in various indoor spaces where people live, stay, and store goods, including elderly care facilities, hospitals, warehouses, and commercial facilities. Traditional reactive approaches, which involve removal and cleaning after mold has already developed, can lead to odors, deterioration of building materials and stored goods, anxiety among users, and increased facility management costs.
In this joint research, the partners aim to build an "Expert Knowledge Fusion AI" that estimates individual mold generation risks. This AI will integrate simple smartphone diagnostics, information on lifestyle habits and facility usage, location data, weather APIs, and over 10,000 cases of on-site knowledge related to consultations, surveys, and construction that Hearts Rich Co., Ltd. has accumulated to date.
Through this initiative, the goal is to shift mold countermeasures from "reacting after occurrence" to "predicting and preventing before occurrence," thereby contributing to improved hygiene management, storage environment control, and better living conditions in residential spaces and facility environments.
The research theme for this joint project is the "Social Implementation of Indoor Mold Risk Prediction and Behavior Modification Systems through the Fusion of Participatory Sensing and Expert Knowledge." The research period is scheduled to run from April 1, 2026, to March 31, 2028.
[Shifting Mold Countermeasures in Living Spaces and Facilities from "Post-Occurrence Reaction" to "Predictive and Preventive Management"]
Mold occurring in indoor spaces is not just an issue of visible dirt and odor; it significantly impacts the quality of life and business operations by causing building material deterioration, quality degradation of stored goods, anxiety about the indoor environment, and burdens on facility management.
However, in ordinary households and facility management settings, it is difficult to determine "which locations are prone to mold," "how much risk the current indoor environment holds," and "whether action should be taken immediately." In many cases, countermeasures are only implemented after visible mold or odors have appeared.
Furthermore, while indoor environment measurement utilizing IoT sensors is an effective method, it faces challenges such as introduction costs, installation efforts, and hurdles to continuous use, making broad adoption in homes and small-to-medium facilities difficult.
Therefore, this joint research will explore a mechanism for predicting mold risk that is more accessible to a broader range of residents and facility managers by combining simple diagnostic information via smartphones, weather data, and the decision logic of experts.
[AI-Powered "Personalized Mold Forecast System" Fusing Smartphone Diagnostics, Weather APIs, and Expert Knowledge]
The "Personalized Mold Forecast System" targeted in this research is an AI-driven prediction tool that estimates individual indoor mold generation risks. It combines indoor environment data, lifestyle habits, and facility usage inputted by users on their smartphones with regional weather data and expert knowledge.
Instead of merely sending alerts like "It is mold season," the system aims to present specific recommendations on which locations require attention and what preventive actions should be taken, based on the space's purpose and usage conditions.
Consequently, residents and facility managers will be empowered to take appropriate actions—such as ventilation, dehumidification, cleaning, improving storage methods, and adjusting inspection frequencies—at the right timing within their daily management routines, thereby proactively suppressing mold generation and recurrence.
[Reflecting Field Knowledge of Mold Countermeasures into the System as an Expert Decision Tree]
Hearts Rich Co., Ltd. is a company that has been working on mold surveys, mold countermeasures, and indoor environment improvements for homes and facilities. Based on the knowledge accumulated through its past on-site responses, the company will organize the conditions prone to mold, location-specific risks, indoor environments where mold recurs easily, and their relationships with usage conditions and management methods. This information will be integrated into the system's decision logic.
In this research, this specialized knowledge will be structured as an "Expert Decision Tree." By combining it with smartphone diagnostics and weather data, the project aims to realize highly practical mold risk assessments grounded in real-world field experience.
[Considering Application to a Wide Range of Indoor Spaces Including Homes, Facilities, and Warehouses]
While this research primarily targets the Japanese living environment, it envisions future application to a broad spectrum of indoor spaces, including elderly care facilities, hospitals, warehouses, and commercial facilities.
The risk of mold generation is influenced by temperature, humidity, building structure, number of occupants, ventilation environments, types of stored goods, and cleaning frequency; thus, the points requiring attention vary depending on the space's purpose. Moving forward, while advancing verifications within Japan, the project will also examine the applicability of mold prediction models factoring in usage purposes and regional differences.
Additionally, the project is considering applications to high-temperature, high-humidity regions such as Southeast Asia.
FAQ
パーソナライズ・カビ予報システムとは何ですか？
スマートフォンの簡易診断、生活環境情報、気象データと、専門家の知見を組み合わせ、AIを活用して室内のカビ発生リスクを個別に予測・提示するシステムです。
この共同研究はどの機関が実施していますか？
慶應義塾大学SFC研究所の環境情報学部中澤・大越研究室（中澤仁教授）と、ハーツリッチ株式会社（代表取締役：穂苅英樹）が共同で実施しています。
共同研究の期間はいつまでですか？
2026年4月1日から2028年3月31日までを予定しています。
IoTセンサーではなくスマートフォンを活用する理由は何ですか？
IoTセンサーは導入コストや設置の手間、継続利用のハードルが高いため、より多くの生活者や施設管理者が手軽に活用できるようスマートフォンを起点としています。
将来的にどのような空間への応用を視野に入れていますか？
日本の住宅環境を主軸としつつ、将来的には高齢者施設、病院、倉庫、商業施設への応用や、東南アジアのような高温多湿地域への展開も視野に入れています。