India’s plagiarism and academic-integrity oversight for higher education runs through a single national regulation: the University Grants Commission’s Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions Regulations, 2018. Unlike the United States, where the Office of Research Integrity (ORI) enforces a federal fabrication-falsification-plagiarism definition tied to Public Health Service funding, UGC’s regulation applies across essentially all Indian higher educational institutions (HEIs) — universities, deemed universities, and colleges — regardless of funding source, and is built around a quantitative similarity-percentage framework rather than a case-by-case misconduct finding alone.
This guide covers what the regulation actually requires: the similarity-percentage tiers and their penalties, the two-panel process a plagiarism complaint moves through, what’s excluded from the similarity calculation, and how Indian institutions are currently handling undisclosed AI-generated content — an area UGC’s 2018 text does not address directly, and where no formal national threshold yet exists.
Legal basis and scope
The regulation was notified by UGC under its statutory rule-making power in 2018 and requires every HEI to establish institutional mechanisms to detect and act on plagiarism in theses, dissertations, and academic/research publications produced by its students and faculty. Institutions are required to run submitted work through anti-plagiarism detection software (commonly Turnitin or iThenticate) and generate a similarity report — what CASRAI’s Dictionary defines more generally as an originality report — before a thesis, dissertation, or manuscript can proceed to evaluation or submission.
The regulation defines plagiarism functionally: the unattributed use of another person’s ideas, words, or other original material, in whole or in part, as though it were one’s own. That framing overlaps closely with the textual plagiarism concept used elsewhere in research-integrity literature, though UGC operationalizes it almost entirely through a numeric similarity index rather than a qualitative test of intent.
The four-tier similarity framework
UGC’s regulation sets four similarity-percentage bands, with different consequences depending on whether the flagged document is a thesis/dissertation or a published academic/research work. In both cases, “Level 0” work is not necessarily plagiarism-free — it simply falls below the threshold the regulation treats as requiring formal action.
Theses and dissertations
- Level 0 — up to 10% similarity: treated as minor/acceptable similarity; no penalty.
- Level 1 — above 10% up to 40%: the student must submit a revised script within a stipulated period, not exceeding six months.
- Level 2 — above 40% up to 60%: the student is debarred from submitting a revised script for one year.
- Level 3 — above 60%: the student’s registration for that programme is cancelled.
Academic and research publications
- Level 0 — up to 10% similarity: minor similarity; no penalty.
- Level 1 — above 10% up to 40%: the author(s) must withdraw the manuscript.
- Level 2 — above 40% up to 60%: withdrawal of the manuscript, denial of one annual increment, and a two-year bar on supervising new master’s or PhD students.
- Level 3 — above 60%: withdrawal of the manuscript, denial of two successive annual increments, and a three-year bar on supervising new master’s or PhD students.
The publication-side penalties are noticeably harsher in career terms than the thesis-side ones at the same percentage band, reflecting that a faculty member’s published output carries institutional accountability (supervision privileges, salary progression) that a student’s thesis resubmission does not.
What’s excluded from the similarity calculation
Because a raw text-match percentage would otherwise flag ordinary scholarly apparatus as “plagiarism,” the regulation excludes certain categories of content before the similarity index is calculated: properly quoted and attributed material, references, bibliography, table of contents, preface, and acknowledgements, along with generic terms, laws, standard symbols, and standard equations. Common-knowledge or coincidental phrasing — sequences of up to roughly fourteen consecutive words that recur across sources for genuinely generic reasons rather than copying — is likewise excluded, provided it isn’t part of a larger matched block. This exclusion logic is one reason a raw similarity-checker score and the regulation’s own similarity index can differ, and why institutions must apply the exclusion rules manually or through checker configuration rather than trusting an out-of-the-box software percentage.
How a case is routed: departmental panel to institutional authority
UGC’s regulation requires every HEI to constitute two tiers of review, and every department within an HEI to notify a first-instance panel. Some secondary summaries of the regulation describe these informally as an “academic misconduct panel” and a “plagiarism disciplinary authority” — the regulation’s own defined terms are the Departmental Academic Integrity Panel (DAIP) and the Institutional Academic Integrity Panel (IAIP), and readers researching this process are more likely to find primary and institutional-policy documents under those names.
- DAIP is constituted at the department level, chaired by the head of the department, with a senior academician from outside the department (nominated by the head of the institution) and a member familiar with anti-plagiarism detection tools. On receiving a complaint or a flagged similarity report, the DAIP investigates and reports its findings and recommendations to the IAIP, typically within a regulation-specified window (commonly cited as 45 days).
- IAIP sits at the institutional level — chaired by a senior academic officer (such as the head of the PG council or equivalent), with a senior academician, an academician from outside the institution, and a member versed in anti-plagiarism software. The IAIP reviews the DAIP’s findings and is the body that actually applies the tiered penalty under the similarity-percentage framework above.
This two-tier structure — a department-level fact-finding panel reporting to an institution-level deciding body — is broadly analogous in shape to the inquiry-then-investigation sequence used in the US ORI research-misconduct investigation process, though UGC’s process is triggered primarily by a quantitative similarity threshold rather than an initial qualitative allegation, and its penalty structure is prescribed by the regulation itself rather than negotiated case-by-case.
AI-generated content: emerging scrutiny, not (yet) a formal threshold
The 2018 regulation predates mainstream generative-AI writing tools and addresses text similarity — copied or paraphrased wording matched against existing sources — not machine authorship as such. UGC has not issued a regulation or amendment formally extending the similarity-percentage framework to undisclosed AI-generated content as of this writing. Readers should treat any claim of an official UGC “AI similarity threshold” with caution unless it cites a specific notified regulation or circular, since none has been identified in UGC’s own published rule-making as of mid-2026.
What is happening, at the institutional rather than national level: a number of Indian universities and technical institutes have been updating their own internal academic-integrity policies since roughly 2024 to treat undisclosed AI-generated text in theses and manuscripts as a form of misconduct in its own right — reasoning that submitting machine-generated prose as one’s own original work misrepresents authorship regardless of whether it also happens to trip the similarity-index threshold. Some of these institutional policies now pair the mandatory similarity report with an AI-content-detection check, or require a disclosure statement describing how generative-AI tools were used (for literature searching, language editing, or drafting assistance) and how the output was verified. This mirrors the disclosure-based approach many journals and publishers have adopted — see CASRAI’s coverage of whether AI-generated text counts as plagiarism and AI paraphrasing tools and plagiarism risk — but it is currently a patchwork of individual institutional policy, not a UGC-wide numeric standard analogous to the 10/40/60% tiers above.
For a research administrator or student in India, the practical implication is: compliance with the UGC 2018 similarity tiers does not automatically mean compliance with your specific institution’s AI-content policy, if it has one — check the institution’s own current guidance rather than assuming UGC’s regulation covers the AI-authorship question.
How this fits alongside other national oversight systems
India’s UGC framework is part of a broader pattern CASRAI tracks across national research-integrity systems: some countries centralize oversight in a single statutory misconduct-investigation body (the US ORI model), others rely on a peer network of institutional offices with a national coordinating or second-opinion function (the Netherlands’ LOWI process, South Korea’s NRF Research Ethics Center and KUCRE), and others build oversight around a specific, quantified compliance mechanism embedded in higher-education regulation rather than a standalone misconduct-investigation agency, which is closer to how UGC’s plagiarism regulation functions. For broader Indian research-governance context, see CASRAI’s coverage of India’s Department of Science and Technology (DST) and the Science and Engineering Research Board (SERB), which handle research funding and international collaboration rather than plagiarism adjudication — UGC’s regulation is specifically a higher-education academic-integrity rule, not a research-funding compliance mechanism. Other national systems in this series include China’s CAST, Pakistan’s HEC, South Africa’s ASSAf/DHET/NRF framework, France’s OFIS, and New Zealand’s Royal Society Code.
Practical takeaways for institutions and researchers
- Run similarity checks well before a submission deadline — a Level 1 finding on a thesis costs up to six months of resubmission time, which is difficult to absorb close to a defense date.
- Apply the regulation’s exclusion rules (quotations, references, generic terms, the roughly fourteen-word common-knowledge allowance) rather than treating a raw software percentage as final — many false “Level 1” flags on legitimate work come from not excluding boilerplate sections properly.
- Understand that DAIP/IAIP is a two-step process with a real investigation stage, not an automatic penalty the moment a percentage is crossed — the DAIP’s findings and recommendation to the IAIP are where individual circumstances get considered.
- Don’t assume UGC’s 2018 regulation resolves AI-authorship questions — check your own institution’s current policy on AI-tool disclosure separately, since this is being set locally, institution by institution, not nationally as of this writing.
- Remember the publication-side penalties (denial of increments, supervision bars) are separate from and can exceed the thesis-side penalties at the same similarity percentage — faculty authors face materially higher stakes than students at equivalent similarity levels.
Frequently asked questions
What is the UGC 2018 plagiarism regulation officially called?
The University Grants Commission (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018.
What similarity percentage is considered “safe” under UGC’s rules?
Up to 10% similarity (Level 0) triggers no penalty under the regulation, though institutions may still expect authors to review and reduce even low-level matches as good practice — a similarity score below 10% is not the same as a certification of originality.
Are DAIP and “Academic Misconduct Panel” the same thing?
They refer to the same function — the first-instance, department-level body that investigates a flagged similarity report — but the regulation’s own defined term is the Departmental Academic Integrity Panel (DAIP), reporting to the Institutional Academic Integrity Panel (IAIP). “Academic misconduct panel” and “plagiarism disciplinary authority” are descriptive phrases used informally by some secondary sources rather than the regulation’s statutory names.
Does UGC have an official AI-plagiarism similarity threshold?
No formal, nationally notified UGC threshold specifically for AI-generated content has been identified as of this writing. The 2018 regulation’s tiers apply to text similarity generally; treatment of undisclosed AI-generated content is currently being handled through individual institutional policy rather than a UGC-wide standard.
Does the similarity check apply to research articles as well as theses?
Yes. UGC’s regulation sets separate (and, at the higher tiers, more career-consequential) penalty tables for academic and research publications in addition to the table governing theses and dissertations.
