From Cybercrime Investigation to Trustworthy AI
An essay on how cybercrime investigation shaped the way I think about evidence, adversarial behavior, and trustworthy AI for high-stakes systems.
AI · Cybersecurity · Learning
Jason Lin
Jason Chia-Sheng Lin · Doctoral researcher at National Yang Ming Chiao Tung University
I study AI and cybersecurity, and share what I learn about speech systems, technical practice, and working across disciplines.
My background in cybercrime investigation shapes how I think about evidence, security, and human judgment.
A few places to start
Selected essays on the questions behind my work.
Designing Speech Evidence Pipelines with ASR and LLMs
A research note on building speech evidence pipelines with ASR, retrieval, and LLMs while keeping outputs grounded in transcript evidence.
From Flat UI to Spatial Interface
Liquid Glass, visionOS, and AI point toward a future where the operating system is redefined less by the kernel than by a new operating surface for search, orchestration, and cross-app work.
Latest writing
Notes, essays, and updates, by publication date.
Recent Work, April-June 2026: Evidence-Aware AI Systems in Practice
A public-safe synthesis of recent work across speech decision stability, clinical workflow support, AI triage demos, realtime voice systems, runtime governance, and AI systems engineering education.
From Flat UI to Spatial Interface
Liquid Glass, visionOS, and AI point toward a future where the operating system is redefined less by the kernel than by a new operating surface for search, orchestration, and cross-app work.
Minimal Disclosure for Fraud Intelligence: Cross-Node Pattern Formation in High-Stakes AI
A research-oriented essay on cross-node fraud intelligence, minimal disclosure, and trustworthy AI design for high-stakes pattern formation under fragmented evidence.
Designing Speech Evidence Pipelines with ASR and LLMs
A research note on building speech evidence pipelines with ASR, retrieval, and LLMs while keeping outputs grounded in transcript evidence.
From Cybercrime Investigation to Trustworthy AI
An essay on how cybercrime investigation shaped the way I think about evidence, adversarial behavior, and trustworthy AI for high-stakes systems.
Follow a thread
Choose a subject and keep reading.
Research & practice
The research, systems, and teaching behind the writing.
CDS-ASR / JANUS
A speech-to-decision stability research line that asks whether downstream decisions remain stable under plausible transcript alternatives.
Project AURA
A local-first meeting-summary workflow that turns corrected transcripts into structured reports while making runtime model readiness and error states visible.
7-day Enterprise Voice AI Systems Onboarding
A Traditional Chinese 7-day consulting-style onboarding tutorial for enterprise voice AI / AI Coach / agent governance delivery, built around architecture, governance, security, deployment, validation, and customer-acceptance evidence.