In the respected halls of higher education, the traditional thesis, the height of independent research and original thought, is faced with an unprecedented challenge. The issue is not of falling standards or apathy, but rather the arrival of a universal, powerful, and invisible collaborator: Generative Artificial Intelligence (AI). The tools of AI are capable of drafting complex arguments, synthesising vast literature, and even generating code; the line between student work and AI-augmented production has blurred into near non-existence.
For thesis supervisors and faculty, this shift is more than a technological inconvenience; it is a profound ethical and pedagogical crisis. The question is no longer if a student used AI, but how a professor can ensure the work demonstrates the critical thinking, research integrity, and original intellectual contribution that define academic rigour.
It can be argued that the era of policing plagiarism is over. To preserve the value of a degree and the integrity of scholarship, higher education must move swiftly to redefine its core supervisory and assessment models, transforming the “Unseen Co-Author” from an adversary into a responsibly managed tool for the next generation of researchers.
The Mandate for Transparency: HEC’s AI Draft Framework
This national challenge is being met by the Higher Education Commission (HEC) in Pakistan, which is in the process of establishing clear boundaries. The HEC's Framework may take a position of cautious engagement, explicitly stating that Generative AI tools cannot be listed as authors and should only be used to improve language and readability. Crucially, the Framework may also mandate that authors must include a declaration statement in their MS/MPhil/PhD theses, detailing the specific AI tool used and the purpose of its application. This approach will, if adopted, shift the burden away from detection alone and towards transparency and accountability.
However, the HEC Framework will also prohibit the use of AI for fundamental research tasks like data gathering, fact generation, or fabrication, which are grounds for academic misconduct. While the HEC is trying to set a clear ethical zero-tolerance policy against using AI to replace original thought, this national mandate will present a dual challenge for supervisors: not only must they check for traditional plagiarism (maintaining the 19% similarity index), but they must now also verify the student’s compliance with the AI use declaration.
The supervisor's role thus will evolve from a plagiarism checker to a “compliance officer” who must evaluate the student’s methodological integrity in an environment where the line between acceptable “language polish” and unacceptable “content creation” is often subjective and tool-dependent.
HEC's Framework vs Global Pedagogical Integration: AI Governance in Academic Research
In contrast, policies emerging from major research universities in technologically advanced states, including China and the US, tend to favour a more decentralised model that emphasises AI literacy, methodological integration, and pedagogical redesign. While universally prohibiting AI from claiming authorship and stressing full human responsibility, these guidelines often treat Generative AI tools as any other research instrument that must be rigorously documented in the Methods section of the thesis.
The true future of global governance in academia lies in a forward-looking pedagogical shift that moves from fearing the AI co-author to meticulously auditing the AI-assisted process, ensuring that human critical thought and accountability remain the undeniable primary source
This approach allows for a broader, justified use of AI in areas like literature review acceleration or data analysis, provided its application is transparent and subject to critical oversight. Furthermore, Chinese universities frequently connect AI usage directly to the central assessment mechanism, warning that an over-reliance on AI will render a thesis indefensible in the viva voce (oral examination). This link between AI use and the core intellectual defence of the research shifts the focus from simple compliance to the demonstration of original critical thought and human value-add — a more sophisticated, yet institutionally varied, form of academic governance.
The Legal and Ethical Abyss: Who Owns the AI-Assisted Thesis?
Beyond the question of academic misconduct lies a deeper legal and ethical vacuum, one that directly impacts researchers in various fields. If Generative AI is considered a “tool” and not an author (as the HEC and global policies agree), the resulting thesis is fraught with unresolved intellectual property (IP) conflicts. Researchers are routinely prompted to upload existing works, published papers, proprietary datasets, or confidential drafts into commercial Large Language Models (LLMs).
This immediate action creates a significant legal risk: Who holds the IP rights to the text and analysis that emerges? Is the student in breach of the publisher’s copyright or the university’s data-sharing agreements? Furthermore, the notorious tendency of LLMs to “hallucinate” false sources and fabricated data must be legally treated not merely as an error, but as a form of machine-mediated academic deception when presented without due diligence. This shifts the supervisor's role yet again, from a compliance officer to a necessary legal and ethical risk manager guiding the student through this new, dangerous digital landscape.
Drawing on the policy comparisons and the identified ethical gaps, a core responsibility now falls directly onto the thesis supervisor. At this stage, it is proposed that their role must evolve from a passive reviewer to an “Integrity Architect.” Supervisors must initiate mandatory, early-stage discussions with their students about the ethical use of AI, moving beyond the simple disclosure required by the HEC. Specifically, they should insist on a “Process Audit Log,” a verifiable record, separate from the thesis itself, that tracks the human-AI interaction, detailing the prompts used, the purpose of the AI assistance (e.g., translation, summary, or syntax check), and the critical human edits applied to the output.
This log serves as the defence of the student's original contribution and intellectual journey. Furthermore, supervisors, particularly those in social and management sciences, must challenge students to critically evaluate the inherent algorithmic bias in their chosen AI tools, ensuring the research compensates for data gaps from the Global South. By embedding this active oversight, the supervisor preserves the integrity of the research while simultaneously training the next generation of scholars to be responsible practitioners in the age of the unseen co-author.
The emergence of Generative AI has not merely presented an integrity challenge; it has fundamentally broken the traditional academic contract, forcing a critical re-evaluation of what constitutes “original work” in the 21st century. While the HEC’s clear boundaries provide essential initial guardrails, the global comparison highlights that genuine research integrity can no longer rely on simple prohibitions or retrospective detection.
The true future of global governance in academia lies in a forward-looking pedagogical shift that moves from fearing the AI co-author to meticulously auditing the AI-assisted process, ensuring that human critical thought and accountability remain the undeniable primary source.
This requires supervisors to embrace their new role as “Integrity Architects” who mandate transparency, scrutinise algorithmic bias, and ensure that a student can vigorously defend not just the output, but the methodological journey recorded in an auditable process log. Ultimately, the quality and credibility of scholarship, whether it concerns the law of algorithms or the fate of the planet, will depend on our collective willingness to embed this new standard of verifiable, human-led intellectual responsibility into the core of every research degree.