Dubai: Fake scientific references and incorrect judicial rulings put university students at risk of academic accountability, which may lead to dismissal and being barred from completing their studies. This comes in light of the spread of artificial intelligence tools capable of producing information and citations that appear reliable, even though some of them may be fabricated or unverifiable.
The danger of what is known as ‘artificial intelligence hallucinations’ lies in the ability of these programs to generate names of studies, authors, references, and judicial rulings that do not exist, or to attribute information to incorrect sources, which may lead students to include them in their research and scientific papers without verifying them, only to find themselves facing violations that affect academic integrity.
In a recent ruling issued by the Dubai Court of Cassation in October, the court upheld the decision to dismiss a master’s student from his university, after he faced accusations of using artificial intelligence to prepare his thesis, in addition to copying materials without referring to their sources. A judicial expert report revealed the presence of incorrect references, citations, and judicial rulings in the thesis, so the case ended with the rejection of his appeal and the upholding of his dismissal from the university, and the rejection of his claim for compensation of Dh100,000.
The ruling came despite the fact that the expert report did not definitively prove that the student used AI techniques. The court affirmed that the lack of conclusive technical evidence proving the use of a specific technique does not mean that the thesis is free from other academic violations, provided that they are proven by sufficient and valid evidence.
Legal experts stressed the importance of distinguishing between using artificial intelligence for linguistic editing and organising ideas, and relying on it to produce information, scientific references, and judicial rulings without referring to their original sources and verifying their accuracy. They emphasised the need to differentiate between unintentional research errors and academic fraud, and to verify the nature and severity of violations before imposing disciplinary penalties that may reach expulsion from the university, while taking into account university regulations and the legal guarantees established for students.
Academic violation
A student’s dream of obtaining a master’s degree in international dispute resolution law turned into a legal dispute with his university, after he passed the exams and reached the final stage of preparing his thesis. He received an email in December 2023 from the academic integrity officer, which included an accusation that the thesis was not prepared with personal effort. An investigation report referred to the conduct committee of the law school in Birmingham, Britain, in February 2024, concluded that he used AI in preparing the thesis, in addition to copying scientific materials without attributing them to their sources.
Following the decision to dismiss him from the university, the student filed an internal appeal which ended with the decision being upheld. He then resorted to the judiciary, demanding the cancellation of the dismissal and the erasure of its effects, and his reinstatement to complete his master’s thesis and obtain his degree, in addition to compensation of Dh100,000 for the material and moral damages that he said he suffered.
During the proceedings, an expert report appointed by the Court of Appeal revealed the presence of incorrect references, citations, and judicial rulings in the letter. The student maintained that the report did not technically confirm his use of AI techniques, nor did it consider the tools used to detect them as conclusive evidence of this, objecting to the reliance on violations related to references and judicial rulings, as they differ from the reason on which his dismissal decision was based.
He also argued that the expert exceeded the limits of his technical task by evaluating the content of the thesis and the extent of adherence to the rules of academic integrity, demanding a distinction between research errors and deliberate cheating, and consideration of the proportionality between the violation and the penalty of dismissal that prevented him from completing his studies.
However, the Court of Cassation in Dubai affirmed that the absence of conclusive technical evidence of the use of a specific artificial intelligence technology does not mean that the thesis is free from other academic violations, as long as they are presented within the scope of the case and are valid for inference, explaining that the court of first instance has the authority to assess the evidence and expert reports, and that the ruling was based on sound reasons sufficient to support the dismissal.
Moreover, objecting to the proportionality of the penalty to the violation does not undermine the validity of the decision, as long as the violation for which university regulations permit the imposition of a penalty has been proven, and the decision has not been proven to violate the rules regulating disciplinary accountability.
Regarding his claim for compensation for the years of study and the fees he incurred, and the damages that he said affected his academic reputation and scientific position, the court affirmed that entitlement to compensation requires proof of the harmful act, the damage, and the causal relationship between them, and that the legality of the dismissal decision and the absence of the harmful act that warrants liability justify the rejection of compensation, and it concluded by rejecting the appeal and obligating the student to pay the expenses and attorney’s fees, with the confiscation of the insurance amount.
Breach of integrity
In the legal context, legal consultant and artificial intelligence governance specialist, Dr. Ammar Ali, confirmed that the use of artificial intelligence in preparing scientific research is not an academic violation in itself, but rather responsibility arises when relying on its outputs without verifying their accuracy, especially if they include fake scientific references or judicial rulings, which constitutes a breach of academic integrity and the credibility of scientific research.
He explained that the UAE legislator emphasised in Article (16), Clause (9), of Federal Decree-Law No. (31) of 2025 on Higher Education and Scientific Research, the necessity for higher education institutions to adhere to the principles of corporate governance, transparency and disclosure, including risk management, ensuring academic integrity and quality of decision-making.
He pointed out that this means that the responsibility of university institutions is not limited to monitoring academic violations and imposing penalties, but extends to setting clear proactive policies and controls that regulate the responsible use of AI tools, define the obligations of students and researchers, and mechanisms for verifying the validity of scientific outputs.
At the international level, Dr. Ali pointed out that the United Nations Educational, Scientific and Cultural Organisation (UNESCO), in its 2023 Guidelines on Generative Artificial Intelligence in Education and Scientific Research, particularly Section (4.2) on protecting the human role, and Section (4.3) on monitoring and verifying generative artificial intelligence systems, called for strengthening human oversight and verifying the reliability of these systems and their suitability for educational and research purposes.
He added that the European Commission, in its 2022 ethical guidelines for the use of artificial intelligence and data in education, identified four basic principles, foremost among them being human efficacy. This means ensuring that humans remain the decision-makers and responsible parties when using artificial intelligence systems, and not relying entirely on their outputs without human review or verification. This preserves the independence of human thought and judgment, especially in preparing and evaluating scientific research and making academic decisions. Along with these are the principles of fairness, humanity, and justified choice, which are linked to the requirements of transparency, accountability, human oversight, and data protection.
Dr. Ali stressed that “governance of artificial intelligence in education does not mean restricting innovation, but rather ensuring its use in a way that preserves academic integrity and enhances human responsibility for scientific outputs.”
Academic governance
Fatima Aboud Al-Mahri, a consultant and international commercial arbitrator specialising in governance, affirmed that governance represents the first line of defence in academic disputes before they reach the courts. She pointed out that clear regulations, defined forms of violations and levels of penalties in educational institutions, and ensuring students are informed beforehand about the limits of using artificial intelligence tools make decisions more accurate under judicial review.
She explained that the case that resulted in the upholding of the master’s student’s dismissal demonstrates the importance of basing disciplinary decisions on clear regulations and documented investigations, and providing avenues for appeal, which strengthens the legitimacy of decisions and protects educational institutions from legal disputes.
She added that academic governance is based on three main dimensions: first, clarity, by setting policies that explicitly define the permitted and prohibited uses of artificial intelligence tools; second, procedural fairness, by enabling the student to defend himself and appeal decisions before a neutral party; and third, proportionality, by determining the penalty according to the severity and nature of the violation.
Al-Mahri warned against relying solely on AI-based detection programs to prove academic misconduct, stressing the need to support their findings with evidence and specialised human analysis.
She stressed that governance is not limited to protecting educational institutions from legal accountability, but extends to protecting students from abuse, preserving the integrity of scientific research, and establishing clear and well-known rules to which everyone can refer when violations occur or disputes arise related to the use of modern technologies in university education.
Justified expulsion
Legal advisor Abdullah Majid Al-Badawi confirmed that a student’s expulsion from university may be justified if they are found to have committed violations that permit such a penalty under university regulations, even without conclusive technical evidence of artificial intelligence use. He pointed out that the Dubai Court of Cassation reached this conclusion in its ruling, stating that expulsion is not merely a matter of an error in the thesis, but rather the nature of the error and the establishment of elements of the violation that justify expulsion according to the applicable regulations.
He clarified that using artificial intelligence in preparing research papers and theses does not, in itself, constitute a violation warranting expulsion, as responsibility depends on the nature of the use and the applicable university regulations. He noted that utilising these tools within permissible limits differs from presenting AI-generated content as a personal effort that violates integrity and disclosure rules, while the researcher remains responsible for the accuracy of the information and references they provide.
Al-Badawi stressed that the presence of incorrect scientific references or judicial rulings in a master’s thesis does not necessarily mean academic fraud, explaining that an oversight in the data of an existing source is different from fabricating a scientific reference or judicial ruling that does not exist, and that distinguishing between the two cases requires examining the seriousness of the errors, their frequency and their impact on the research, and listening to the student’s explanation of them, according to the definition of the violation in the university regulations.
Regarding the reliance on AI-generated content detection programs, Al-Badawi stressed that the results of these programs are technical indicators that require examination and verification, and do not become conclusive evidence simply because they are issued. He pointed out the need to balance them with the content of the research, its sources, drafts, and the student’s statements, especially since the Court of Cassation’s ruling did not grant the detection tools conclusive authority, but rather relied on the evidence that the lower court was satisfied with regarding academic violations, despite not being certain about the use of the technology.
He pointed out that the student’s reaching the final stage of the master’s program, passing the exams, and paying the tuition fees does not prevent the university from making a decision to dismiss him, whenever a violation is proven that allows the regulations to impose this penalty, and the accountability procedures have been completed. He explained that the ruling did not stipulate that the lesser penalties must be exhausted before dismissal in all cases, but rather it refers to the applicable regulations, taking into account the proportionality between the violation and the penalty, the integrity of the procedures, and the guarantees of defence.
Al-Badawi called on higher education institutions to develop clearer regulations that define the permitted and prohibited uses of artificial intelligence tools, the requirements for disclosing their use, mechanisms for verifying the validity of references and information, and procedures for investigation, penalties, and grievances, in order to protect students from accountability resulting from the ambiguity of academic controls.
He stressed that these requirements are consistent with the legislative framework regulating higher education in the country, noting that Article (16), in clauses (9) and (10), of Federal Decree-Law No. (31) of 2025 on Higher Education and Scientific Research, combines ensuring academic integrity with employing modern technologies.
Source: Emirates 24|7

