Is a High AI Score Considered Academic Misconduct? How Detection Flags Affect Degrees in Chinese Higher Education

An elevated AI detection score on an academic paper is an automated screening signal generated by an algorithm rather than a formal institutional finding of academic misconduct; whether a score affects defense scheduling or requires revision depends on the current rules of the degree-granting unit. Furthermore, institutional notices—such as Shandong Jianzhu University's spring 2026 graduate notice, which represents one university notice, not a national threshold—explicitly prohibit artificially lowering the proportion of AI-generated content through deletion or tampering, treating verified falsification under academic integrity regulations.

HumanPen Team

· 8 min read

Is a high AI rate considered academic misconduct, and does it affect your degree?

An elevated AI detection score is an automated screening signal generated by an algorithm, not a formal institutional finding of academic misconduct; its specific impact on thesis defense and degree progression depends on the current rules of each degree-granting unit rather than an automatic nationwide rule. Furthermore, institutional notices (such as that of Shandong Jianzhu University) strictly prohibit artificially lowering the proportion through deletion or tampering, treating verified falsification under academic integrity regulations.

When an academic thesis produces an elevated AI score on a screening system, that percentage is an automated screening output and not by itself a formal institutional finding of academic misconduct. In terms of institutional and legal standards, an automated screening percentage calculated by software is fundamentally distinct from a formal institutional finding of academic misconduct. The governing statutory framework, including the Academic Degree Law of the People's Republic of China, does not establish a specific software metric as a uniform national threshold for wrongdoing; formal institutional determinations rely on established investigation and resolution procedures rather than an automated software report.

However, this distinction does not mean students can disregard an elevated score. What happens after an elevated score is reported depends entirely on what the reader's degree-granting unit's current notice specifies. If an institutional inquiry confirms serious integrity violations—such as generative AI ghostwriting or deliberate falsification—the degree-granting unit may take formal disciplinary measures under its regulations and the statutory framework. Similarly, attempting to circumvent screening through text deletion or tampering introduces independent compliance risks. Readers must therefore verify what their own institution's current notice specifies.

Screening signal vs. institutional finding: how universities assess AI detection

Much of the anxiety surrounding thesis AI screening stems from conflating preliminary software indicators with substantive institutional verdicts. In university administration, these operate at fundamentally different levels:

```mermaid flowchart TD A["Automated Screening Signal<br/>(Automated screening output from screening tool)"] --> B["Institutional Review<br/>(Handled by degree-granting unit under current notices and rules)"] B --> C["Compliant with Institutional Rules / Permitted to Proceed"] B -->|Suspected Ghostwriting or Fraudulent Evasion| D["Institutional Inquiry<br/>(Conducted by degree-granting unit under its procedures)"] D --> E["Institutional Determination<br/>(Handled under university rules and statutory degree procedures)"] ```

1. Automated Screening Signal (机检初筛信号)

The percentage output by a screening tool is an automated screening output.

  • Nature of the Output: The percentage is an automated screening output and not by itself a formal institutional finding of academic misconduct.
  • Administrative Role: The score functions as a preliminary screening reference established by administrative departments under local institutional rules, not as an automated legal determination of wrongdoing.

2. Institutional Misconduct Finding (学术不端实质认定)

Academic misconduct is a formal disciplinary conclusion carrying serious administrative and legal effects, governed by institutional academic integrity rules and statutory legislation such as the Academic Degree Law of the People's Republic of China.

  • Authority Held by Degree-Granting Units: Academic misconduct determinations are made by the degree-granting unit through established procedures. Software tools possess no independent legal or administrative authority to declare misconduct.
  • Inquiry Centers on Independent Work: Investigations focus on whether the academic work was independently produced by the author. Contemporaneous materials that students may independently preserve (such as draft version histories, experimental records, or notes) are simply records students can preserve for their own reference, not statutory legal requirements or guaranteed proof.

The hidden risk: why artificially manipulating scores is treated as misconduct

When confronted with an elevated preliminary flag, some students explore techniques such as deleting flagged paragraphs, disrupting sentence structures, or mechanical synonym replacement to push the percentage down. Public university policy records show that such deliberate evasion strategies present direct institutional compliance risks.

In its official notice titled Notice on Doing a Good Job in AIGC Detection of Graduate Dissertations in the Spring Semester of 2026 (关于做好2026年春季学期研究生学位论文AIGC检测的通知, published March 30, 2026), the Graduate School of Shandong Jianzhu University established explicit rules regarding thesis process management:

“导师应加强学位申请人的学位论文写作过程管理,严禁在形式上弄虚作假,严禁以删除、篡改等手段人为降低人工智能生成的内容比例。” (Advisors shall strengthen the management of the degree applicant's thesis writing process; it is strictly prohibited to engage in form-based falsification, and strictly prohibited to artificially lower the proportion of artificial intelligence generated content through means such as deletion or tampering.)

The notice immediately defines the institutional consequence:

“一经发现与核实,学校将按学术不端相关规定给予严肃处理。” (Once discovered and verified, the university will handle the matter seriously in accordance with relevant regulations on academic misconduct.)

Precise scope and significance of this notice

  • Institutional Scope: This provision comes from a specific administrative notice issued by the Graduate School of Shandong Jianzhu University for the spring semester of 2026 (`https://www.sdjzu.edu.cn/yjsc/info/1055/3131.htm`). It is one university notice, not a national threshold.
  • Regulatory Focus: The notice explicitly characterizes "artificially lowering the proportion of artificial intelligence generated content through means such as deletion or tampering" as form-based falsification (在形式上弄虚作假). While a preliminary automated score must be addressed through established institutional procedures, intentionally tampering with text to deceive screening systems is classified as falsification that triggers formal academic misconduct procedures.
  • Implication for Candidates: Mechanically manipulating text merely to suppress a score counter risks degrading academic clarity and coherence, and transforms what is essentially a technical reference check into a violation of academic integrity rules. Students must never employ deceptive methods to evade screening.

How detection results affect degree progression: governed by institutional rules

While an elevated AI screening score is not an automatic revocation of a degree or an immediate disciplinary penalty, its effect on thesis progression is determined entirely by the current rules of each degree-granting unit:

1. Institutional workflows vary without a uniform national automated hold

There is no nationwide, automatic rule dictating that exceeding a given score automatically halts blind review (盲审) or bars defense registration across all Chinese universities. The effect is whatever the reader's degree-granting unit's current notice specifies. For example, under Shandong Jianzhu University's spring 2026 notice—which represents one university notice, not a national threshold—screening was conducted on a sample basis (covering 20% of defense applicants) with an advisory warning threshold of 40% (the proportion of intelligently generated content in principle not exceeding 40%). Manuscripts exceeding 40% triggered an early warning requiring the applicant to revise the thesis to lower the proportion and resubmit it to their college. This demonstrates that screening protocols belong to local institutional discretion; readers cannot assume that one university's notice reflects universal national practice.

2. Substantive reviews focus on independent authorship

Where an institution's rules call for human evaluation of a flagged manuscript, the fundamental inquiry remains whether the core scholarship was independently conducted and authored by the candidate. Students can preserve chronological records of their work (such as draft revisions, tracked changes, or research notes); these are records candidates can preserve for their own reference, not statutory legal requirements or guaranteed proof, and their evaluation depends entirely on the degree-granting unit's policies.

3. Statutory framework of Academic Degree Law Article 37

At the statutory level, Article 37 of the Academic Degree Law of the People's Republic of China (effective January 1, 2025) establishes the overarching statutory framework for degree denial and revocation: where a degree applicant or degree recipient engages in academic misconduct such as plagiarism, forgery, or tampering, or obtains a degree by fraud—including having a thesis written by another person or using artificial intelligence to ghostwrite a dissertation—the degree-granting unit may, by resolution of its degree evaluation committee (经学位评定委员会决议), decide not to grant the degree or to revoke an already conferred degree. The statute establishes severe legal outcomes through the formal committee resolution process to address substantive fraud and ghostwriting, rather than prescribing an automated software score as an automatic statutory penalty.

What students must verify: consulting your degree-granting unit's current rules

Because policies on AI-assisted tools vary across institutions and disciplines, candidates facing thesis screening must not rely on online rumors and should verify current requirements through official institutional channels:

  • [ ] 1. Consult current official directives from your Graduate School or Academic Affairs Office Check your university's official portal for the thesis screening notice issued for the current semester. Confirm whether AIGC detection is conducted, whether it applies to all students or a sample, whether advisory thresholds exist, and what revision or review workflows are specified for flagged papers. The actual effect depends entirely on what the current notice specifies.
  • [ ] 2. Confirm specific college and departmental implementation guidelines The acceptable boundaries for language polishing, grammar correction, or conceptual assistance differ across disciplines; verify the permitted scope with your department or advisor.
  • [ ] 3. Use only officially designated institutional screening platforms Avoid uploading manuscripts to unverified third-party websites, which provide unreliable metrics and expose unpublished academic work to intellectual property risks.
  • [ ] 4. Strictly refrain from deletion, tampering, or commercial lowering services Do not attempt to manipulate manuscripts through arbitrary phrase replacement, character corruption, or commercial "AI lowering" vendors. As stated in Shandong Jianzhu University's notice, artificially lowering proportions through deletion or tampering constitutes form-based falsification handled under academic integrity regulations.
  • [ ] 5. Maintain chronological records of your research and drafting process Cultivate rigorous archiving habits throughout your academic work: retain draft versions, tracked changes, research notes, and laboratory or survey records. These are records students can preserve for their own reference, not statutory legal requirements or guaranteed proof, but they serve as personal reference materials documenting your drafting trajectory.

Sources and references

The institutional notices and statutory provisions cited in this article are derived from the following public sources:

  1. Graduate School, Shandong Jianzhu University (山东建筑大学研究生处): Notice on Doing a Good Job in AIGC Detection of Graduate Dissertations in the Spring Semester of 2026 (关于做好2026年春季学期研究生学位论文AIGC检测的通知, published March 30, 2026). Provisions on dissertation process management, the 20% sample screening, the 40% advisory warning threshold, revision procedures, and the explicit prohibition against form-based falsification and artificially lowering AI-generated content proportions through deletion or tampering. URL: `https://www.sdjzu.edu.cn/yjsc/info/1055/3131.htm`
  2. Ministry of Education of the People's Republic of China (教育部政策法规): Academic Degree Law of the People's Republic of China (中华人民共和国学位法, adopted April 26, 2024, effective January 1, 2025). Article 37 (statutory provisions governing refusal to award or revocation of degrees by resolution of the degree evaluation committee in cases of academic misconduct, AI ghostwriting, or fraud). URL: `https://www.moe.gov.cn/jyb_sjzl/sjzl_zcfg/zcfg_jyfl/202404/t20240426_1127804.html`

KEEP READING