Academic Integrity Platform

Part of the worlds-biggest-software-project initiative.

A unified platform for plagiarism detection, AI writing detection, and citation verification -- built to support fair, evidence-based academic integrity enforcement.

Academic Integrity Platform combines text similarity analysis, AI-generated content detection, and citation hallucination verification in a single system with integrated case management. It is designed for higher education institutions that need to enforce academic honesty policies consistently and fairly, without relying on fragmented toolchains or substituting opaque automated scores for human judgment.

Why Academic Integrity Platform?

No single product covers the full integrity lifecycle. Institutions currently stitch together separate tools for detection (Turnitin, Copyleaks, GPTZero) and adjudication (Maxient), with no data continuity between them. Detection reports must be manually exported and re-attached to conduct cases.
AI detection is sold as a costly add-on. Turnitin charges separately for AI writing detection on top of its core similarity product, creating budget friction that leaves some institutions without AI detection capability at all.
Citation hallucination detection does not exist in commercial platforms. AI tools routinely generate fabricated references. Open-source tools like CheckIfExist and RefChecker can verify citations against bibliographic databases, but no commercial integrity platform integrates this capability.
False-positive rates fall unevenly on vulnerable students. Research consistently shows that non-native English speakers are disproportionately flagged by AI detectors. Current tools provide scores but no contextual tooling to help adjudicators assess whether a flag reflects genuine misconduct or structural bias.
Privacy-first alternatives are in demand. Turnitin's practice of retaining student submissions in its proprietary database has triggered FERPA and GDPR concerns. Some European institutions have restricted or banned its use entirely due to data sovereignty issues.

Key Features

Detection Engine

AI-generated content detection with sentence-level granularity and confidence intervals, covering current major LLM families (GPT-4/5, Claude, Gemini, Llama, DeepSeek)
Text similarity detection against web content, academic journals, and institutional submission archives with source-match highlighting and percentage scoring
Citation hallucination detection via bibliographic reference verification against CrossRef, Semantic Scholar, and OpenAlex
Code plagiarism and AI-generated code detection for computer science assignments
Multilingual detection support

LMS Integration

LTI 1.3 (LTI Advantage) integration with Canvas, Moodle, Blackboard, D2L Brightspace, and Schoology
Grade passback via Assignment and Grade Services
Submission workflow embedded within the student and instructor LMS experience, requiring no additional accounts

Case Management and Adjudication

Structured workflow from incident report through evidence packaging, student notification, appeal tracking, and outcome recording
Audit trail connecting detection reports to case records and adjudication outcomes
Configurable workflow stages aligned to institutional academic integrity policies
Exportable evidence chains for accreditation and legal review

Fairness and False-Positive Mitigation

False-positive contextualisation tools highlighting known bias factors: non-native English writing patterns, standard academic phrases, and discipline-specific vocabulary
Spectrum classification (fully human / lightly AI-assisted / moderately AI-assisted / fully AI-generated) rather than binary scoring
Adjudicator guidance linking detection thresholds to policy-consistent actions

Formative Student Tools

Student-facing self-check mode providing similarity, citation quality, and AI feedback before final submission
Formative feedback framing that supports learning rather than serving purely as a penalty mechanism
Instructor-configurable visibility settings for student access to their own reports

AI-Native Advantage

The platform uses AI not just for detection but to improve the fairness and consistency of the entire integrity process. False-positive risk scoring contextualises flags against known bias patterns rather than leaving adjudicators to interpret raw scores without guidance. Citation integrity analysis goes beyond checking whether a reference exists to verify whether cited sources actually support the claims attributed to them (semantic claim-source alignment). An assignment design advisor analyses assignment prompts and suggests modifications that reduce AI-completion feasibility -- an upstream intervention that complements downstream detection.

Tech Stack & Deployment

The platform targets self-hosted and cloud deployment models, with configurable data residency to address GDPR and institutional data sovereignty requirements. LMS integration uses the LTI 1.3 standard with Assignment and Grade Services for grade passback. Citation verification integrates with CrossRef, Semantic Scholar, and OpenAlex APIs. The system requires FERPA-compliant submission storage with role-based access controls and SOC 2-level security assurance. SIS integration (Banner, PeopleSoft, Workday) is planned for pulling live student demographic and enrolment data into case records.

Market Context

The academic integrity tools market was estimated at approximately $1.1 billion in 2024 and is growing at roughly 14% CAGR, driven by the surge in AI writing tool adoption since 2022. The market serves approximately 20,000 higher education institutions globally. Turnitin's dominant position is being contested by bundled-detection alternatives (Copyleaks) and AI-native startups (GPTZero, Pangram, Proofademic), while institutional pricing varies widely -- The Markup (2025) documented significant price variation for identical Turnitin products across institutions.

Project Status

This project is in the research and specification phase. Contributions, feedback, and domain expertise are welcome.

Contributing

We welcome contributions from developers, domain experts, and potential users. See CONTRIBUTING.md for guidelines.

Important: All contributions must be your own original work or clearly attributed open-source material with a compatible licence. Copyright infringement and licence violations will not be tolerated and will result in immediate removal of the offending contribution. If you are unsure whether a piece of code, text, or other material is safe to contribute, open an issue and ask before submitting.

Licence

Licence to be determined. See discussion for context.