ChillStack Launches 'AI Safety Solution' to Address AI Risks via Education, Diagnosis, and Defense for Secure AI Development and Operations
ChillStack has launched the 'AI Safety Solution', offering a comprehensive approach of education, diagnosis, and real-time defense to protect enterprise AI applications from risks like prompt injections and hallucinations.
📋 Article Processing Timeline
- 📰 Published: April 23, 2026 at 22:06
- 🔍 Collected: April 23, 2026 at 13:31
- 🤖 AI Analyzed: April 23, 2026 at 23:35 (10h 3m after Collected)
ChillStack Inc. (Headquarters: Shibuya-ku, Tokyo, CEO: Michiaki Ito, hereafter "ChillStack"), which supports the challenges of government agencies and major enterprises with AI and security, announces the launch of the "AI Safety Solution" that comprehensively assists in grasping and countering security risks in AI-equipped services.
This solution supports risk mitigation when companies implement AI in their services, covering the development, implementation, and operation phases. By combining the three services of "Education," "Diagnosis," and "Defense" provided by ChillStack, companies can identify AI-specific risks in advance and build a framework to appropriately counter them, realizing an environment where they can safely accelerate their AI business.
Background
Since the emergence of ChatGPT in 2022, generative AI has become an indispensable technology for promoting DX in enterprises. Many SIers and business companies are proceeding with AI implementation in their services. According to a survey by ChillStack (*), it was found that 56% of companies are already providing and operating systems incorporating AI.
On the other hand, "AI-specific risks" such as attack methods like prompt injection and hallucinations have materialized, becoming serious concerns that threaten system quality and reliability. Especially in large-scale development sites and quality control departments, they face challenges such as "accuracy of risk assessment varies by project" and "traditional quality assurance processes cannot objectively prove the safety of AI." Formulating organizational evaluation standards and establishing effective defense mechanisms during the operation phase are urgent tasks.
To solve these issues, ChillStack is launching the "AI Safety Solution" to comprehensively support everything from understanding risks to implementing countermeasures. By standardizing risk assessments and implementing advanced security measures, we support the secure development and operation of AI-equipped services in enterprises.
*Source: ChillStack "AI Governance Fact-finding Survey (December 2025)"
About the "AI Safety Solution"
Targeting the entire supply chain of "Development," "Construction/Implementation," and "Deployment/Operation" of AI-equipped services, the "AI Safety Solution" provides comprehensive risk countermeasures in each phase by combining three services: "Education," "Diagnosis," and "Defense."
Education: Secure AI Training
A training service jointly provided with Mitsui Bussan Secure Direction, Inc., where participants can acquire the knowledge and technology to safely proceed from the design and planning of AI-equipped services to their operation. We offer a wide range of courses, from overviews for management to advanced hands-on training for engineers, supporting the improvement of security awareness and technical skills across the organization.
Details here: https://jpsec.ai/
Diagnosis: AI Diagnosis
Consultants and white-hat hackers proficient in AI conduct multifaceted investigations and diagnoses of AI-specific attack methods like prompt injections and system configuration flaws, in addition to traditional vulnerabilities. We identify specific risks such as information leakage and hijacking in advance and propose effective countermeasures.
Details here: https://pentest.chillstack.com/
Defense: Safia (AI Protection System)
A service that monitors employee inputs and outputs to AI in real time and automatically blocks inappropriate use. Since detection models can be flexibly customized according to use cases, it provides practical defense, such as preventing the leakage of personal and confidential information and blocking unauthorized access to systems.
Details here: https://safia.chillstack.com/
*With the launch of the "AI Safety Solution," "Stena AI" has been renamed to "Safia."
About ChillStack Inc.
Company Overview
Company Name: ChillStack Inc.
Location: Yoyogi Air Heights 206, 5-16-10 Sendagaya, Shibuya-ku, Tokyo
Founded: November 2018
CEO: Michiaki Ito
Corporate Website: https://chillstack.com/
Business Description:
"Supporting social innovation with cutting-edge AI and security technologies."
ChillStack provides world-class AI security technology solutions that comprehensively resolve risks associated with the advancement of AI and DX. For enterprises, we deploy services that support fraud/anomaly analysis and safe AI utilization. In collaboration with government agencies, we handle even more complex...