Elith Inc. (Representative Director, CEO & CTO: Koki Inoue; Bunkyo-ku, Tokyo; hereinafter "Elith") is pleased to announce that Representative Director, CEO & CTO Koki Inoue will speak at "AI × Security Study Session #9," a study session hosted by CDLE Cyber Security, a group within CDLE, the AI community of the Japan Deep Learning Association (JDLA).
▼Details and registration:https://jdla.connpass.com/event/396941/
With the emergence of the latest AI models such as Claude Mythos and Fable, cybersecurity is reaching a major turning point on both the attack and defense fronts. At this event, two experts at the forefront of AI × Security will explain the risks these models pose and how to counter them, drawing on their own verification results.
▼Event overview
| Item | Details |
|---|---|
| Date and time | Thursday, June 25, 2026, 20:00–21:00 |
| Format | Online (Zoom) |
| Organizer | CDLE Cyber Security (an official community of the Japan Deep Learning Association) |
| Participation fee | Free |
| Registration URL | https://jdla.connpass.com/event/396941/ |
▼Timetable
| Time | Content | Speaker |
|---|---|---|
| 20:00–20:05 | Opening | Yuki Kawada |
| 20:05–20:20 | The new phase of cybersecurity revealed by Claude Mythos | Yuki Kawada |
| 20:20–20:50 | AI threats and security measures in the Mythos era | Koki Inoue |
| 20:50–20:55 | Q&A | All participants |
| 20:55–21:00 | Closing | akio ueki |
Speaker profiles
"The new phase of cybersecurity revealed by Claude Mythos" | Yuki Kawada
Joined NTT DATA Financial Technology Corporation in 2007 and worked on the development and operation of various financial systems. Later shifted focus to AI-driven fraud detection (AI for Security), handling MLOps and R&D. In recent years has also worked on security infrastructure design for large-scale systems and on researching Security for AI.
Starting from Mythos and Fable, which have drawn much attention recently, this talk explains the new phase in the relationship between AI and cybersecurity, incorporating the speaker's own verification results from vulnerability assessments.
"AI threats and security measures in the Mythos era" | Koki Inoue (Representative Director, CEO & CTO, Elith Inc.)
Earned a master's degree at the Japan Advanced Institute of Science and Technology (JAIST), researching materials exploration for quantum computers. Currently enrolled in the doctoral program at the Graduate School of Medicine, Tohoku University, also researching medical AI in the field of medical physics. Active across research, development, and real-world deployment, including books and contributions on generative AI and LLMs, and presentations at international conferences.
From an AI security perspective, this talk clearly explains the specific threats posed by Mythos and how to counter them, incorporating the latest trends.
About CDLE Cyber Security
CDLE (Community of Deep Learning Evangelists) is Japan's largest AI community, made up of people who have passed the G Certification and E Certification administered by the Japan Deep Learning Association (JDLA). CDLE Cyber Security works to bridge the knowledge gap between security engineers and AI engineers and to promote the fusion of AI and security.