The Miilk

AI Life-Story Interview Cuts Senior Cognitive Test Refusal to 8%

reporter-profile
Genie Kim 2026.08.31 16:39 PDT
AI Life-Story Interview Cuts Senior Cognitive Test Refusal to 8%

Lebengrida's IWAS returns a cognitive report from the same session that produces a senior's life storybook; general availability October 2026

SEOUL, South Korea, Aug. 31, 2026 — Older adults refuse cognitive testing far less often when nobody calls it a test: a South Korean AI service that records seniors' life stories now returns a cognitive risk score from the same conversation, cutting refusal from 42% to 8%.

7.4 Million Cases Hide Behind a Refused Test

About 22% of U.S. adults 65 and older live with mild cognitive impairment, according to a 2022 JAMA Neurology study. University of Southern California researchers reviewing Medicare records for 40 million beneficiaries found clinicians had diagnosed only about 8% of expected cases, leaving 7.4 million undetected. MCI converts to dementia at 10% to 15% a year. Paper instruments such as the MMSE take the form of an exam, and older adults avoid exams.

IWAS never had a test to remove. A synthetic child avatar that lip-syncs in real time asks about hometown, childhood and family, building each question from the previous answer. One tap starts the session; the interview runs on speech alone. Lebengrida calls this self-referential memory assessment: the screening reads a person's own life story instead of questions with right answers. That interview was already running in homes and care facilities. What is new is what Lebengrida reads from it.

Monthly Reports Build a Baseline Before October Launch

Users complete one session a week, 12 minutes each against 30 for a paper instrument. Every month returns two files: a storybook narrated in the user's own voice, and a cognitive report. Twelve months produce an annual trend against the user's own baseline and a forward projection.

"We already had the conversations," said Hyunjoo Kang, CEO of Lebengrida. "What we learned is that a person's own memories are the most honest test material there is. Nothing to prepare for, nothing to fail."

The screening version is running in a proof of concept with a South Korean insurer's VIP membership and Lebengrida has set general availability for October 2026. Pilot participants described the sessions as putting their lives in order, not taking a test. Lebengrida holds registered Korean patents covering IWAS and has filed international application PCT/KR2024/017522.

About Lebengrida

Lebengrida develops reminiscence-based cognitive education and screening for older adults. Its IWAS service records AI-led life-story interviews that produce narrated storybooks and mild cognitive impairment risk reports; BrainUp provides follow-on cognitive training. Lebengrida serves care facilities, insurers and families across South Korea. Founded in 2018 and based in Busan, South Korea. Learn more at lebengrida.co.kr.

Media Contact

Hyunju Kang

lebengrida@naver.com

이 기사와 관련있는 기사 현재 기사와 관련된 기사들 입니다.

In the AI Era, the Way I Work Has Completely Changed

"It's already been over 4 to 5 months since I last used the design tool Figma. Eighty percent of my daily routine is spent directly in Claude Code, prototyping and testing."The AI innovation that hit Silicon Valley developers is rapidly spreading into adjacent areas. A major shift is happening in the world of design, as code writing is now entrusted to AI coding agents, while humans mainly take on judgment and verification roles.The changes in work styles brought about by AI adoption first affected software developers. As code-writing speed increased rapidly and quality improved, entrusting coding to AI became the mainstream. According to Google CEO Sundar Pichai, even within Google, where the world's top developers gather, AI agents write 75% of new code.👉 "75% of new codes generated by AI"... The future shown by Google's 'Gemini Agent Platform'The key point is that the AI-driven work revolution is not limited to the development domain. In Silicon Valley, non-developers are also using AI to improve work efficiency or turn their ideas into reality through AI coding, making it a daily routine. Designers who work closely with developers are prime examples.

박원익 2026.07.01

Goover by Saltlux: Revolutionizing AI Search and Research

Since OpenAI launched ChatGPT on November 30, 2022, the world of generative AI has exploded. Tools like Claude, CoPilot, Perplexity, MidJourney, and Firefly have reshaped industries from writing and translation to coding and design.With AI innovation surging, the question for individuals and businesses is clear: how can generative AI enhance personal competitiveness and organizational success?“AI must solve real-world problems to create real value,” says Kyungil Lee, CEO of Saltlux, a leading generative AI company based in South Korea. Interview with TheMiilk during a visit to the U.S.,Lee stressed that AI isn’t just a playground for experimentation. “AI literacy is about much more than using prompts or tools—it’s about defining clear problems and leveraging AI to solve them,” he explained.AI Literacy: Problem-Solving in PracticeTrue AI literacy, according to Lee, starts with the ability to identify and articulate challenges. “Being able to ask specific questions and communicate those to an LLM (large language model) is critical,” he says. Lee points to SK Group’s AI disruption initiative as a model for applying AI to real-world challenges. “Their approach of defining practical problems and solving them with AI creates valuable use cases that others can learn from.”The key, Lee believes, lies in accumulating problem-solving experience. “This process not only sharpens your understanding of AI but also maximizes its effectiveness,” he said.A New Player in AI SearchSaltlux, under Lee’s leadership, is taking a step forward in the generative AI space with its latest offering, Goover, a search and research platform. Introduced in June through Saltlux’s U.S. subsidiary, the tool is designed to provide what Lee calls a 'Re-Search', combining traditional search with advanced AI insights. Available in both Korean and English, Goober’s service has also expanded to mobile platforms.What sets Goover apart? According to Lee, it’s the combination of Saltlux’s proprietary LLM, called LUXIA, and its established expertise in vector search technology. Goover navigates multilingual web information and even paid subscription services, presenting users with precise answers and trustworthy sources.“In our own tests, Goober achieved a 98% accuracy rate and a 97% recency rate in Korean search queries,” Lee said, noting that its performance rivals that of ChatGPT and surpasses Perplexity in some areas. Beyond search, Goover offers features like auto-generated reports and personalized content recommendations, aligning with the growing demand for hyper-personalized digital experiences.The Future of Generative AIDuring a visit to Goover’s Silicon Valley HQ, Lee reflected on the future of the AI industry. While generative AI tools continue to evolve at breakneck speed, the Saltlux CEO emphasized that their ultimate success will hinge on how effectively they address real-world challenges.“AI tools are no longer just helpful—they’re becoming essential,” Lee said. “The organizations and individuals who can combine problem-solving skills with AI capabilities will lead in this era of disruption.”As companies like Saltlux push the boundaries of what generative AI can achieve, the challenge for users is clear: to harness these tools not just for exploration but for meaningful impact. Whether through search, personalization, or industry-specific applications, AI’s transformative potential is only beginning to unfold.The following is the full interview between TheMiilk reporter Wonik Park and Saltlux Guber CEO Kyungil Lee.

박원익 2024.11.17