cv.zekkx.icu

Shibo Yan（厳 世博）

厳 世博 · ゲン セイハク

~ $ cat role.txt

grad student · tokyo session 00:00:00 scroll ↓
01 About 01 / 05

About Me

Based in Tokyo, with a background in automation engineering and systems work. Hands-on with server operations, AI tooling, and quantitative analysis — now studying what corporate AI disclosure actually reflects.

For my master's thesis I built a six-year panel of 115 listed Japanese information & communication firms from EDINET filings, combining text analysis, panel regression, and a close reading of financial statements — from data pipeline to interpretation.

Technical background, research mindset — curious about where technology meets accounting.
02 Skills 02 / 05

Skills

Skill matrix
Skill Level Evidence
Data & Analysis
Statistical modelingProficientMCM/ICM 2020 · Master's thesis
Panel regressionProficientMaster's thesis (FE / RE / CRE)
SPSS · MATLABWorkingMCM/ICM 2020
PythonLearningResearch data processing
Accounting & Finance
BookkeepingWorkingBookkeeping Grade 2 (簿記2級)
Financial statement analysis · Management accounting · Corporate financeWorkingGraduate coursework
Languages
ChineseNative—
JapaneseBusinessJLPT N1
EnglishWorkingCET-4 · 525
03 Projects 03 / 05

Projects

  1. [01] IEEE Conference Paper · 2026 Co-first Author

    A Regulatory Placebo: The Failure of Mandatory GenAI Labeling

    Jingyi Chen†, Chaofan Bu†, Xuesong Li*, Shibo Yan†

    Compares mandatory labeling of AI-generated content across five jurisdictions and argues it works as a regulatory placebo — hard to enforce, technically self-defeating, and weakly justified. Proposes content- and risk-based governance instead.

    • GenAI Governance
    • Watermarking
    • Comparative Law
  2. [02] Master's Thesis

    Does AI Disclosure Reflect Substance?

    Tests whether AI mentions in the annual securities reports of 115 Japanese information & communication firms reflect real transformation. Over FY2020–FY2025, mentions grew mostly as generic wording, never predicted future operating ROA, and rose within firms after weak years — most consistent with symbolic signaling.

    • Textual Analysis
    • Panel Data
    • Disclosure Research
  3. [03] MCM/ICM 2020 · Problem E

    A Safe Global Target for Single-Use Plastic

    Built a six-compartment mass-flow model of the global plastic system. At 2016 infrastructure only about 113 of 138 Mt/yr of single-use plastic waste could be handled safely; keeping the environmental stock under a 250 Mt ceiling implies a 2050 target of 11.9–19.6 Mt/yr — an 87–92% cut — delivered through an equity-first roadmap.

    • Mass-flow Model
    • Sensitivity Analysis
    • Environmental Policy
04 Education 04 / 05

Education

  1. 2025 — 2027

    Meiji UniversityM.Acc.

    School of Professional Accountancy · Accounting

    Tokyo · Japan

    Thesis: does AI-related disclosure reflect substance? Evidence from Japan's information & communication firms

    progress --%

  2. 2018 — 2022

    Southwest UniversityB.Eng. + B.A.

    Control Engineering + Japanese Studies

    Chongqing · China

05 Contact 05 / 05

Get in Touch

Email
shibo_yan@outlook.jp
ORCID
0009-0003-5447-7744
Institution
Meiji University · Graduate School of Professional Accountancy
Location
Surugadai Campus · Chiyoda-ku, Tokyo
Status
M.S. 2nd Year · Expected March 2027