Plagiarism and similarity reduction in Hyderabad

What similarity limit does each Hyderabad university set for a PhD thesis?

Anything from under 10% to 24%. Who checks it differs just as much. Every row comes from the institution’s own document.

Treat your university’s figure as a ceiling, not a target. The same file gives different numbers on Turnitin and DrillBit, and with different exclusion settings. Our national plagiarism limits guide explains the UGC 2018 levels and exclusions. This page sticks to the local rules those levels don’t cover.

Why does JNTUH’s 3% single-source rule catch theses that pass the 24% limit?

Because there are two limits, not one. Section 11 of JNTUH’s doctoral regulations in force since 2021-22 sets 24% overall. It also allows no more than 3% from a single source other than your own publications. The second limit is the one that bites. Picture a literature review that leans on one long review article. Or a methods chapter adapted from a single standard paper. Either can put 5% or 6% on one source while the thesis total looks comfortable. So when we read a JNTUH report, we sort the source list by percentage first. Anything near 3% gets dealt with before we touch the total.

The university’s application for anti-plagiarism check fixes the process:

• You apply only after the Departmental Research Committee rates your colloquium satisfactory.
• The thesis goes in as a single Word file on a CD. The CD is labelled with your name, hall ticket number, supervisor, co-supervisor, department and year of admission. The regulations ask you to bring it to R&D in person.
• The colloquium report and a no-dues certificate go with it. Your supervisors sign that they have reviewed the thesis and will abide by the university’s plagiarism rules.
• If the figure is over the limit, you have three chances in all. Each revised copy is due within a month, on the resubmission form. A fourth needs a genuine reason and the Vice-Chancellor’s approval.

Those one-month windows sit inside the six months JNTUH allows between colloquium and submission. It’s far less work to check chapters before the colloquium than to rewrite after a failed check. The JNTUH format rules are in our JNTU thesis format guide.

How does the University of Hyderabad count your own published papers, and which tool does it use?

Most of the University of Hyderabad’s plagiarism guidance is about the scholar’s own papers. First, the limit. A Controller of Examinations circular of 9 January 2017 set the cap for PhD, M.Phil. and M.Tech. theses at 15%, effective from 1 February 2017. The version reproduced in the 2026-27 prospectus keeps the same wording, but it sets the cap at 10%. So if seniors submitted under the 15% rule, their advice is out of date.

The rest of the guidelines:

• If you’re first author of a published paper, its similarity with your thesis is ignored in the overall figure.
• If someone else led the paper, the question goes to your supervisor together with the Head or Dean, one paper at a time.
• Each paper travels with the thesis as an annexure. That’s either the paper in print, or its abstract with the acceptance on journal letterhead or official email. A chapter based on a paper may note where it appeared.
• Theses should not be a reproduction of papers. Supervisors are asked to encourage students to rewrite them.

The check itself sits with the library. Your declaration in the thesis submission formats encloses “a report on plagiarism statistics from the University Librarian”. Which tool does the library use? Its own pages say the IGM Library subscribes to Turnitin. It ran a Turnitin similarity and AI tool refresher on 30 July 2025, and it held a DrillBit demonstration on 17 June 2025. So ask the library which tool it will use for your thesis. If it’s Turnitin, our guide to reading a Turnitin report helps. For DrillBit, see our DrillBit guide.

What do Osmania’s 20% AI-content limit and IIT Hyderabad’s AI declaration mean for your drafts?

Osmania is the only Hyderabad university we found with a number for AI. Clause 28.1 of the 2026-27 rules puts the thesis through the university’s approved software. The similarity index must not be above 10%, and AI-generated content similarity must not be above 20%. Clause 27.4(a) makes you sign an undertaking with both figures. The UGC’s 2018 regulations say nothing about AI, so the 20% is Osmania’s own. What the rules leave out matters too. They don’t name the detector, and they don’t say which pages it reads, or what happens between 20% and a rejection. Clause 28.3 does say that a submission missing any required document is rejected. And clause 26.6 lets a result be withheld or cancelled for plagiarism even after it is published. Our Osmania thesis format guide describes the rest of the Osmania submission file.

IIT Hyderabad goes the declaration route instead. On its AI declaration form, you certify that any AI-assisted content was critically reviewed, verified and acknowledged. Its similarity index declaration asks each Doctoral Committee member to record the report as acceptable or unacceptable.

AI detectors estimate probability. They don’t prove authorship, and they’re known to flag plain, formulaic academic English written by people. Say an Osmania report flags text you wrote yourself. Your drafts, notes and version history are then your evidence, so keep them from the first chapter. Our guide to AI tools in a PhD sets out what’s safe to use. Our own edits are human edits, in tracked changes. We don’t run text through paraphrasing or “humanising” tools. That’s misconduct, and it’s also a reliable way to raise an AI score.

EFLU and MANUU. At EFLU the PhD handbook uses the same software on coursework. A term paper, project report or research proposal over 10% gets an F grade, and you resubmit within 15 days. At MANUU, the 2023 regulations make the library’s Plagiarism Verification Certificate part of the academic clearance you need before submission.

Working with us from Hyderabad

You send the similarity report and the document, and we look at both before quoting. Then, on a video call, we go through the report with you. Which matches are properly cited quotations? Which fall under the standard exclusions, and which are citation slips? And which passages need rewriting? Corrections come back to you as tracked edits in Word. With them comes a new report that names its checker and settings.

Your university sets the timing. A JNTUH scholar gets one month between checks at R&D. So the useful moment for us is before the colloquium, chapter by chapter. That’s when a single-source match over 3% can be found and fixed with no deadline pressing. At the University of Hyderabad, we mark which matches come from your first-author papers. The Librarian’s count ignores them, so we focus on the rest. An Osmania scholar faces two figures, so we look at the similarity matches, and also at any passages you drafted with AI help that need rewriting in your own words. At EFLU your supervisor runs the check, so we aim to have the thesis clean before it reaches the supervisor.

Some things stay with you. That means delivering the CD to JNTUH’s R&D, and collecting the Librarian’s report at UoH or the Verification Certificate at MANUU. It means every signature on the forms, too. We can’t submit on your behalf, and we don’t use any university’s account. Our ethics policy sets out why each change is one you make and can explain.