No Student Should Be Penalised Based On AI Detection Tool Alone: CJI Surya Kant
| The CJI pointed out that many AI-detection tools often returned false results.
Chief Justice of India (CJI) Surya Kant has cautioned law schools against relying on artificial intelligence (AI) detection software to establish academic misconduct, warning that unreliable tools could wrongly brand honest students as cheats.
"No student should therefore be held guilty of misconduct on the word of a machine alone," he said while delivering the inaugural address at the Commonwealth Legal Education Association's (CLEA) First Annual International Conference on Legal Education, 2026, in Hyderabad on Saturday.
The CJI observed that many institutions now rely on software that claims to determine whether an essay or research paper was written by a student or generated by a machine. However, he cautioned that such software was far from reliable.
Referring to a 2023 study published in the journal Patterns, titled "GPT Detectors Are Biased Against Non-Native English Writers", he noted that seven widely used AI detectors wrongly flagged more than three out of every five essays written by non-native English speakers as machine-generated.
The study found that while the detectors performed almost flawlessly with essays written by schoolchildren in the United States, they frequently misidentified essays written by non-native English speakers. It tested 91 TOEFL essays and 88 essays written by students in the United States.
Highlighting the implications for Commonwealth countries, where many students write in English as a second or third language, the CJI warned:
"In a Commonwealth where so many students write in English as a second or third language, an honest student could thus be branded a cheat on the strength of a number that no one can explain, and that student cannot meaningfully challenge."
He cautioned against treating the verdict of such software as proof of academic misconduct.
AI Should Assist, Not Replace Students' Independent Thinking
Addressing the broader challenges posed by generative AI to legal education, the CJI said that these tools had unsettled the traditional assumption that an examination paper or research assignment reflected what a student actually knew.
He noted that generative AI could now summarise judgments, suggest arguments and produce polished answers within minutes.
"The question, therefore, is no longer whether students will use these tools, but whether the work they submit still reflects their own reasoning, understanding, and appreciation for the law," he said.
The CJI said that an outright ban on AI tools would be neither practicable nor wise, as students would eventually enter a legal profession that was already using them.
Referring to the approach adopted by the Supreme Court of Singapore, he noted that court users are permitted to employ generative AI tools but remain fully responsible for everything they file.
He cautioned that banning such tools would only drive their use underground, where it could neither be guided nor examined. Instead, law schools should establish clear rules and teach students responsible use of technology.
Law Schools Must Rethink Assessment Methods
The CJI said that the better safeguard against misuse of AI was to establish clear guidelines before an assessment begins, rather than attempting to identify misconduct after academic work had been submitted.
Students must know which uses of AI are permitted, which require disclosure and which would amount to handing over the very thinking that an assessment is intended to test, he said.
He also suggested that assessment methods should evolve to ensure that students genuinely understand and can defend the work they submit.
Written assignments could be supplemented with oral explanations, a defence of the legal authorities relied upon, or exercises requiring students to identify errors in machine-generated answers.
"In the end, a law school must be able to certify, with confidence, what its graduates themselves know and can do," he said.