University of Reading AI Detection Policy: Turned Off, Not Activated, or Unmentioned?

*All three descriptions exist in the University of Reading's published documents, but they belong to three distinct documents published across three different academic years. In November 2023 (Version 1.0), the university stated that Turnitin's AI detection facility was "turned off." In June 2025 (Version 1.4), the wording shifted to "we have chosen not to activate this functionality yet," explaining that the institution needed confidence in its reliability, an understanding of its limitations, and formal guidelines before enabling it. In August 2026, the university approved a new overarching policy statement, Preparing Students for an AI-Enabled World (Version 1.0), which contains zero mentions of Turnitin or AI detection. However, the university's live assessment guidance page continues to instruct markers that student work must not be uploaded into AI detection tools under an interim approach, and links directly to the June 2025 Version 1.4 document.*

While the three documents are tabulated in our overview of uploading your draft to an AI detector, the apparent discrepancy across these texts is not a contradiction in current enforcement, but a failure to check document titles, dates, and version identifiers. Below is what each official document actually states, how the university's written framing evolved, and why citing an older, unlinked PDF can backfire during an academic integrity review.

The widely repeated quote that Reading has "turned off" AI detection comes from a specific historical document: Academic Integrity and Generative AI Tools, marked on its cover as `(V1.0 NOV 2023)`.

In that November 2023 guidance, the university's Centre for Quality Support and Development (CQSD) gave staff an unambiguous operational update:

"Turnitin's facility to detect AI use is turned off due its low reliability and the risk of false accusations of misconduct."

The same document instructed staff on alternative handling:

"Staff should be aware of the signs of AI misuse but should not submit students' work into GAIT for checking, on the grounds we do not have their permission to do this, and these checkers are not reliable."

That document reflected the sector's initial reaction following Turnitin's April 2023 launch of its AI writing detector. However, as institutional guidance matured, Reading revised its staff documentation.

In June 2025, CQSD released an expanded 17-page document titled Generative AI Tools and Assessment, marked on the cover as `V1.4 (JUNE 2025)`. In that document, the phrasing changed noticeably:

"Turnitin released AIWDC (Artificial Intelligence Writing Detection Capabilities) functionality in April 2023. At the University of Reading, we have chosen not to activate this functionality yet."

The document immediately explained the institutional prerequisites for changing that position:

"Before doing so, we need to be confident in its reliability and effectiveness, understand its limitations, and establish appropriate policies and guidelines for its use."

It added an observation on detection evasion:

"Early use in the sector has shown that while Turnitin AIWDC can reasonably detect simple copy-pastes from basic AI prompts, its effectiveness decreases when common methods to evade AI detection are applied."

The difference between the 2023 and 2025 documents is subtle but critical. Saying a tool is "turned off due to low reliability" describes an active decision to disable an operational switch. Saying "we have chosen not to activate this functionality yet" frames detection as an unadopted vendor capability that remains under evaluation pending institutional confidence and policy frameworks.

In August 2026, the University Board for Teaching, Learning and Student Experience (UBTLSE) approved a new institutional policy statement: Preparing Students for an AI-Enabled World: University Policy Statement on Artificial Intelligence in Teaching, Learning and Assessment.

The document control block at the end of the statement records its official status:

• Version: `Version 1.0`
• Approving authority: `UBTLSE`
• Approval date: `12/08/26`
• Start date: `12/08/26`
• Next review: `May 2027`

If readers review this 6-page institutional statement expecting a definitive decree on Turnitin or detection software, they will find nothing. The words "Turnitin," "detector," and "detection" appear zero times throughout the entire document.

Instead, under the heading "People first. AI second.", the August 2026 policy articulates high-level educational principles:

"Our approach starts with people and educational purpose, not technology. AI should support human learning, creativity and achievement rather than become an end in itself."

On the integrity of assessment, the statement establishes broad institutional duties rather than technical tool mandates:

"The University will maintain the academic standards of its awards and ensure that assessment remains a valid and reliable means of evaluating student achievement."

It immediately pairs that commitment with student accountability:

"Students remain responsible for the work they submit, regardless of any use of AI technologies."

Does the omission of detection tools in this policy mean Reading abandoned its restrictions? No. The university maintains operational boundaries on its teaching guidance website. On Reading's live CQSD web page titled GenAI in Assessment and Feedback (retrieved September 2026), the university provides instructions to markers under its marking and feedback guidance:

"Student work must not be uploaded into GenAI tools or AI detection tools under the current interim approach."

Under its practical boundaries list for staff, the page repeats:

"Student work should not be uploaded into AI detection tools."

Regarding the tool platforms used by staff, the page immediately clarifies:

"Staff should use institutionally supported tools and follow DTS and IMPS guidance."

Crucially, when that same live page invites educators to "Read the fuller GenAI tools and assessment guidance," the hyperlink points directly to the June 2025 Generative AI Tools and Assessment PDF (`V1.4`). The university approved its overarching institutional statement in August 2026 while continuing to link its June 2025 guidance document on assessment tools.

When an institution releases a major policy update that omits a controversial topic, readers often interpret the silence as a covert policy change.

In university administration, treating omission as action is a fundamental mistake. Institutional documentation functions in layers:

1. Strategic policy statements (such as Preparing Students for an AI-Enabled World) establish ethical principles, pedagogical direction, and board-level commitments. They rarely catalogue specific vendor add-ons or software switches.
2. Operational staff guidance (such as the CQSD GenAI in Assessment and Feedback page and the Version 1.4 PDF) defines day-to-day administrative rules, handling procedures, and interim software boundaries.

When the August 2026 policy omitted Turnitin, it did not secretly activate AI detection across the university, nor did it formalize a permanent ban. It simply kept strategic policy separate from operational assessment procedures.

Inferring that "because Turnitin is not named in the 2026 policy, the university must now be using it" is unsubstantiated. Equally, assuming that the absence of a detection policy gives students free rein or guarantees that work will never be examined for authenticity contradicts the explicit statement that "students remain responsible for the work they submit, regardless of any use of AI technologies."

When students face an inquiry regarding alleged AI use, a common defense tactic is to search online, locate an official university document, and quote it to demonstrate that the institution does not recognize detection tools.

If a student at Reading—or any university with evolving documentation—quotes an old PDF without auditing its version, that defense can collapse upon inspection.

Consider what happens when someone cites the November 2023 PDF:

1. Current pages no longer link to the document: The November 2023 PDF (`academic-integrity-and-genai.pdf`) is no longer linked from Reading's active CQSD pages, though the direct URL remains accessible.
2. Later documents use different wording: If an investigating officer or panel looks up the reference, they will see that the June 2025 document (Generative AI Tools and Assessment) uses "chosen not to activate this functionality yet," while the August 2026 overarching policy statement contains no mention of detection tools at all.
3. The university explicitly warns against caching stale copies: On Reading's main CQSD artificial intelligence page, the university includes a clear disclaimer: > "This page is regularly updated. Please revisit this page when seeking advice and avoid making local copies that may become out of date."
4. Credibility is damaged: Presenting a 2023 quote from an unlinked document as current institutional policy suggests that the student did not consult live guidance, shifting the panel's focus away from verifiable evidence of their writing process.

In an academic integrity review, panels evaluate whether work is authentically the student's own, focusing on evidence of authorship rather than quotes from unlinked historical documents.

To avoid relying on obsolete guidance—whether at Reading or any other institution—students and researchers can follow a systematic four-step audit before citing any institutional policy document:

1. Locate the document control table: Look at the cover page, header, or final page for formal metadata. Identify the exact version number (such as `Version 1.0` vs. `V1.4`), the approving authority (e.g., `UBTLSE` or Senate), the approval date, and the scheduled next review date. If a PDF lacks document control metadata entirely, treat its temporal scope with caution.
2. Trace the parent navigation path: Do not rely on search engine landing pages. Navigate from the university's current academic year index (such as the active teaching support or academic regulations hub) and confirm that the document is actively linked from the live menu. At Reading, for instance, navigating from the CQSD Artificial Intelligence hub to "GenAI in Assessment and Feedback" leads to the active link for the June 2025 document (Generative AI Tools and Assessment), whereas the November 2023 PDF has no link from these live pages. If a PDF can only be found via direct search queries and has no inbound links from the current site structure, it is likely an unlinked historical file.
3. Track phrasing shifts across revisions: When comparing multi-year documents, watch for shifts in modality and operational verbs. A transition from "turned off" to "not activated yet," or from mandatory prohibitions to case-by-case tutor discretion, signals that institutional thinking has evolved. Cite the most recent revision and quote the exact wording.
4. Distinguish high-level strategy from interim procedures: Recognize the administrative role of the document you are reading. An overarching board statement on teaching philosophy will rarely address specific tool features. Check whether procedural boundaries are maintained in separate, subordinate guidance documents (such as assessment handbooks or interim faculty FAQs).

By applying this audit method, you evaluate institutional rules as they actually exist in the current academic session, rather than relying on historical artifacts found across the web.

• University of Reading, Centre for Quality Support and Development (CQSD). Artificial Intelligence portal page. Retrieved September 18, 2026. https://www.reading.ac.uk/cqsd/artificial-intelligence
• University of Reading, University Board for Teaching, Learning and Student Experience (UBTLSE). Preparing Students for an AI-Enabled World: University Policy Statement on Artificial Intelligence in Teaching, Learning and Assessment (Version 1.0, approved August 12, 2026). https://www.reading.ac.uk/cqsd/-/media/project/functions/cqsd/documents/ade/Preparing-Students-for-an-AI-Enabled-World.pdf
• University of Reading, CQSD. GenAI in Assessment and Feedback guidance page. Retrieved September 18, 2026. https://www.reading.ac.uk/cqsd/artificial-intelligence/assessment_and_feedback
• University of Reading, CQSD. Generative AI Tools and Assessment (Version 1.4, June 2025). https://www.reading.ac.uk/cqsd/-/media/project/functions/cqsd/documents/ade/genai-tools-and-assessment.pdf?la=en&hash=BF3295266A4A91A8933E802108C48A67
• University of Reading, CQSD. Academic Integrity and Generative AI Tools (Version 1.0, November 2023). https://www.reading.ac.uk/cqsd/-/media/project/functions/cqsd/documents/ade/academic-integrity-and-genai.pdf