Unifa Revises AI Policy to Protect Children's Authenticity, Safety, and Security

Unifa Inc. (headquartered in Chiyoda-ku, Tokyo; CEO: Yasuyuki Toki, hereinafter "Unifa"), a company leveraging Caregiving AI™ technology to solve social challenges in childcare and parenting, announces the revision of its "AI Policy" in August 2026. This revision aims to continue providing better services through its comprehensive childcare ICT platform "Lookmee," and to further pursue an environment where AI can be used safely and securely.

1. Background and Purpose of the AI Policy Revision

While AI technology adoption is advancing across society, the widespread use of generative AI has made it easy for anyone to create images, leading to the emergence of "false and misleading information" through deepfakes as a new social risk. According to the Ministry of Internal Affairs and Communications' report "ICT Literacy in Data 2026," 80.6% of internet and social media users feel they may have encountered false or misleading information※1. Eliminating the objective risk of misidentifying AI-generated content and ensuring high governance and ethical standards among AI providers will become increasingly important※2.

The core value of Unifa's photo and documentation services, including "Lookmee Photo," lies in recording "the authentic facts" of children's growth. This AI policy revision aims to establish a shared understanding both internally and externally that "children's appearances will not be generated or altered" across all communications—including service development, advertising, public relations materials, and presentation documents—so that customers can use "Lookmee" with even greater peace of mind and safety.

※1: "AI Provider Guidelines (1.2 Edition)" (Ministry of Internal Affairs and Communications, Ministry of Economy, Trade and Industry, revised March 2026) ※2: "ICT Literacy in Data 2026" (Ministry of Internal Affairs and Communications, May 2026)

Four Key Commitments in the Revised "AI Policy" (Areas Where AI Will Not Be Used)

Unifa defines both proactive AI applications and areas where AI will not be used, establishing the following four principles to protect the factual integrity and dignity of children:

① We will not generate or alter children's appearances

We will not use AI that could output non-existent or factually inaccurate depictions or scenes of children.

Creating new depictions of children (including generating images, videos, or audio impersonating real children)

Altering children's appearances in photos or videos to differ from reality (e.g., swapping facial expressions, adding or removing people, compositing, beautification)

Using AI-generated expressions in Unifa's services, advertisements, PR materials, or presentation documents that could be mistaken for real children Note: Adjustments such as color balance, brightness, noise reduction, cropping, and blurring—corrections that do not alter the recorded facts—are not included. The standard is "we perform corrections that do not change facts, but we do not perform edits that alter facts."

② We will not use photos or data entrusted to us for training AI that could generate children's appearances

Photos, videos, audio, and other personal data entrusted to us will not be used to train or improve external providers' generative AI models. When using external generative AI services for transcription, summarization, or formatting, we will adopt contracts and settings that ensure input data is not used for the external provider's AI model training or improvement※.

※In-house training and evaluation may be conducted within the scope of our privacy policy, with appropriate security measures, to improve the accuracy of identification and analysis functions such as facial recognition and photo blur detection developed and managed by Unifa.

③ We will not evaluate or determine children's development using AI alone

We will not make definitive judgments about children's development or potential based solely on AI. AI-based analysis and suggestions are positioned as support for professional human judgment by caregivers and others. Human verification and judgment will always be included in evaluations concerning children.

④ We will not use AI without disclosing its involvement

When providing outputs where AI has been involved in generation or analysis to parents and childcare facilities, we will clearly indicate AI's involvement. We will not provide such outputs in a way that could be mistaken for human-created content without disclosing AI use.

Unifa Inc. AI Policy Details

The full text of the revised AI policy can be viewed at: https://unifa-e.com/company/ai-policy.html

Unifa will comply with the revised AI policy, protect children's "authentic selves" and safety, and continue developing and providing services toward realizing a "child-centered society" where all children can thrive and everyone involved in childcare can work with smiles.

2. About Lookmee and Unifa

What is the Comprehensive Childcare ICT "Lookmee®"? <https://lookmee.jp/>

Childcare professionals handle a wide range of tasks, including managing children's arrival and departure, monitoring naps, creating childcare diaries and plans, preparing documents for parents and municipalities, and managing staff schedules.

The "Lookmee" series automatically aggregates data such as arrival/departure records, temperature checks, meals, and toileting, along with photos taken via Lookmee Photo and sleep data automatically recorded by Lookmee Nap Check. This data is then automatically transferred to forms and communication logs. With Lookmee Nap Check, sensors detect body movement, and the app automatically records body position. By digitizing childcare operations (DX), the "Lookmee" series significantly reduces workload, creating mental and time relief for caregivers. The saved time allows caregivers to spend more quality time with children. The photo documentation feature fosters an environment where caregivers can easily share insights about children's growth, enhancing rich communication. These efforts contribute to increasing caregiver fulfillment and improving childcare quality. To date, the "Lookmee" series has been introduced in over 20,000 cases cumulatively and adopted by approximately 70 municipalities※.

※As of September 2024 (based on internal research)

Unifa Inc. <https://unifa-e.com/>

Unifa is a startup in the "Childcare-Tech" field, aiming to solve social challenges related to childcare and parenting. With the purpose of "creating new social infrastructure for family happiness worldwide," we develop and provide the comprehensive childcare ICT service "Lookmee" using cutting-edge technology centered on Caregiving AI™. Through our "Smart Kindergarten®, Smart Nursery School®, Smart Childcare Center®" vision, we contribute to building a rich society where people can work while raising children. Since being selected as the inaugural champion of the Startup World Cup in 2017, we have won multiple prestigious awards. We were selected as a J-Startup in 2021 and as a J-Startup Impact in October 2023. We also serve as the representative director of the Japan Child DX Promotion Association, established in February 2023. In August 2025, we will participate in METI's "Japan Star