Interdisciplinary Researcher · Taiwan

RZVN

ZON RZVN

Explore Research → Contact

Current Public Focus

AI context, conversational contextual risk, and user-side safety

This site is the public entrance to a broader interdisciplinary research program. The current published branch focuses on how AI dialogue contexts accumulate consequence for the user over time, and how those consequences can be named, assessed, and bounded without collapsing into clinical or legal overclaim.

User-side contextual hallucination is one branch phenomenon within that larger AI context program. The current public branch moves from contextual structure, to branch-level problem definition, to post-interaction assessment, and then to bounded technical detection.

A conversation can remain technically acceptable while becoming contextually unsafe for the user.
Start With the Essay → View Research Stack

Current Public Branch

The published branch is built as a linked sequence rather than a group of isolated outputs.

Research Stack Map
  • CXC-7 and CXOD-7 define the contextual structure and operation layer.
  • USCH isolates user-side contextual hallucination as one branch phenomenon within the broader AI context program.
  • USCI and A-CSM move from post-interaction assessment to bounded detection.

Why This Matters Now

Public Use Has Already Arrived
Before Safety Language Caught Up

The problem is not theoretical timing. The public evidence already shows emotionally consequential use at scale.

72%

Nearly three in four teens had used AI companions at least once, according to Common Sense Media.

Common Sense Media

64%

Pew reported that 64% of U.S. teens used AI chatbots, with roughly three in ten using them daily.

Pew Research

1 in 8

Brown University reported that about one in eight adolescents and young adults use AI chatbots for mental health advice.

Brown University

2025 → 2026

The EU AI Act, OMB M-25-21, and Singapore’s agentic AI framework show governance acceleration, while user-side assessment remains comparatively thin.

EU AI Act · OMB M-25-21

Research Spine

The current public branch moves from contextual structure and contextual operation to a branch phenomenon, post-interaction assessment, and a bounded system release.

Research Stack Map

The Research Program Is Structured,
Not Fragmented

AI context is the umbrella problem. CXC-7 and CXOD-7 establish the structural language. USCH defines one branch phenomenon in that space. USCI proposes post-interaction assessment. A-CSM turns the public branch into a reproducible release boundary for technical inspection.

The work is cumulative. Each layer narrows the problem without collapsing it into a single claim, and each release keeps a different boundary between explanation, formalization, and executable inspection.

Research Architecture Methods Formal Sources

Articles

Essays and Briefings Built From the
Published Research Branch

These pieces turn the formal work into readable arguments, source-linked briefings, and public-facing explanations without drifting away from the papers.

Open Article Index → Open Press Room

A-CSM

The Executable Layer of the
Current Public Branch

A-CSM is not the whole research program. It is the bounded implementation layer that detects conversational contextual risk after interaction artifacts are available. The public-safe core exists for reproducible inspection, not for clinical or legal decision-making.

Open A-CSM → Validation Boundary

Press and Institutions

A Press Room Built Around
Reportable Themes

The press room now organizes the current public work around reportable themes, current public context, approved framing, and article-level briefings rather than a short media bio alone.

Open Press Room → Press Contact