Dubai Court Upholds Student's Dismissal Over AI-Generated Thesis With Fabricated Rulings

The court issued its ruling on October 29. It rejected the student's appeal and his claim for Dh100,000 in compensation, ordered him to pay legal costs and Dh2,000 in lawyers' fees, and ruled that he would lose his appeal security deposit.

The student was enrolled in a master's programme in international dispute resolution law. According to the report, he had finished his coursework and examinations and reached the thesis stage when the university began disciplinary proceedings. On December 7, 2023, the university's academic integrity officer emailed him to say the thesis was not his own work.

An investigation report dated February 8, 2024, written by an investigator from Birmingham Law School for the Faculty Conduct Committee, found that he had used artificial intelligence to generate parts of the thesis and had copied material without properly crediting the sources. The committee dismissed him from the programme, and his internal appeal was rejected.

How the Case Moved Through the Courts

The student then sued. He wanted to be reinstated, allowed to finish and submit his thesis, and awarded the degree, along with Dh100,000 for material and moral damages. The Court of First Instance dismissed the case on February 10, 2025. The Court of Appeal appointed an expert and upheld the dismissal on July 30, 2026.

In his final appeal, the student argued that AI-detection software is unreliable and cannot prove misconduct. He also said the court-appointed expert had gone beyond the assignment by examining the thesis content and its legal references, and that the university should have imposed a lighter penalty.

The Court of Cassation disagreed on each point. It held that concerns about the reliability of AI-detection tools did not invalidate the other evidence. That evidence included the expert's findings of fabricated or inaccurate legal references and of judicial rulings that either did not exist or had been misrepresented. The court said checking the thesis content and verifying its references was relevant to judging the integrity of the work. On compensation, it said a claim requires proof of wrongdoing, actual harm and a direct link between the two. Because the university's decision was lawful, the student had not shown any wrongdoing.

Why It Matters

The ruling deals with a problem universities and employers are increasingly running into: detection software that is widely seen as unreliable. In this case the court did not treat a detector's verdict as the deciding factor. It relied on something that can be checked directly, namely citations to cases and sources that do not exist. Such fabricated references are a known failure of generative AI tools, and lawyers in several countries have already been sanctioned for filing them.

Dr Ammar Ali, a legal consultant specialising in AI governance, told Gulf News that using AI in research is not in itself academic misconduct. He said the real question is whether researchers take responsibility for checking what the tools produce. He pointed to Article 16, Clause 9 of Federal Decree-Law No. 31 of 2025 on Higher Education and Scientific Research, which requires UAE higher education institutions to uphold governance, transparency and disclosure, including protecting academic integrity.

Ali said universities should go beyond disciplining violations and publish clear policies on the responsible use of AI. Those policies should set out what students and researchers are obliged to do and how AI-generated content will be checked, with fair procedures and penalties proportionate to the offence. He also cited UNESCO's 2023 guidance on generative AI in education and research, which stresses human oversight and validation of AI systems.

For universities, and for any organisation that relies on written work, the case suggests that the burden of verification lies with the person submitting the document. Checking sources may also hold up better as evidence than AI-detection scores.