Banning AI tools neither practical nor wise: CJI
Chief Justice of India Surya Kant said banning AI tools outright would be neither practical nor wise. He noted that the technology has already become part of professional and academic life. Legal education, he argued, needs to address how students use AI rather than act as if it can be removed. The key distinction is whether AI supports a student’s work or takes over the student’s thinking. Surya Kant called for clear rules and for students to understand that difference. He also warned that reliance on AI should not weaken academic integrity, independent reasoning or the ability to apply legal principles. He proposed that written assessments be supported by oral explanations, defences of legal authorities and tasks that ask students to find errors in AI-generated answers. He also urged CLEA to develop model guidelines for disclosing AI assistance in legal research. These steps would let law schools address AI use while continuing to assess students’ own understanding.
What did the Chief Justice of India say about banning AI tools in legal education?
Chief Justice of India Surya Kant said banning AI tools outright would be neither practical nor wise. He noted that the technology has already become part of professional and academic life. Legal education, he argued, needs to address how students use AI rather than act as if it can be removed.
The key distinction is whether AI supports a student’s work or takes over the student’s thinking. Surya Kant called for clear rules and for students to understand that difference. He also warned that reliance on AI should not weaken academic integrity, independent reasoning or the ability to apply legal principles.
He proposed that written assessments be supported by oral explanations, defences of legal authorities and tasks that ask students to find errors in AI-generated answers. He also urged CLEA to develop model guidelines for disclosing AI assistance in legal research. These steps would let law schools address AI use while continuing to assess students’ own understanding.
What is the difference between using AI as an aid and letting it replace a student's own thinking?
Using AI as an aid means a student uses the technology to support learning or legal work while remaining responsible for understanding and judging the result. Letting it replace their thinking means relying on its output instead of developing their own reasoning and ability to apply legal principles. That difference matters because legal education is meant to build skills, not just produce a finished answer.
The article gives examples of ways to test those skills. Students could explain their answers orally, defend the legal authorities they rely on, or identify mistakes in AI-generated responses. These tasks require them to show what they understand, rather than simply submit text.
Surya Kant said banning AI completely would be impractical because it is already part of academic and professional life. The priority, he said, should be clear rules and students’ understanding of AI’s proper role. Supreme Court judge PS Narasimha also warned that excessive dependence could weaken independent thought, reasoning and sound judgment.
How often did the study find AI detectors wrongly identified essays by non-native English speakers as machine-written?
The study Surya Kant referred to found that widely used AI detectors wrongly identified more than three in every five essays written by non-native English speakers as machine-generated. That is a high rate of false identification in the group described. It matters because a detector’s result may be treated as evidence about a student’s work, even when the result is wrong.
The article does not name the study or give further details about how it was conducted. It reports the finding as part of Surya Kant’s concern about institutions’ growing reliance on software designed to detect AI-generated content. He warned that an honest student could be branded a cheat based on a machine’s finding.
His proposed safeguard was clear: no student should be found guilty of misconduct on a machine’s word alone. He suggested supplementing written assessments with oral explanations and other ways to assess understanding. The finding supports his call for careful rules rather than treating detector output as conclusive proof.
What could happen to a student if a school treats an AI detector's result as proof of cheating?
If a school treats an AI detector’s result as proof, a student whose own essay is wrongly flagged could be accused of cheating or branded a cheat. Surya Kant raised this risk after citing a study in which detectors wrongly identified more than three in every five essays by non-native English speakers as machine-generated.
The article does not describe a specific student case or state what punishment a school might impose. It does make clear that detector findings can be wrong and that an honest student may face an unjust accusation if a machine’s result is accepted without further checks. A detector result, on its own, does not establish that a student committed misconduct.
Surya Kant said no student should be held guilty on a machine’s word alone. He recommended supplementing written assessments with oral explanations, defences of legal authorities and exercises that ask students to find errors in AI answers. Those checks can give educators more evidence of what a student understands.
How could oral explanations or asking students to find errors in AI answers help assess their understanding?
Oral explanations and error-spotting tasks can help educators assess whether students understand the law behind their work. Instead of judging only a written submission, a teacher can hear a student explain their reasoning, defend the legal authorities they used, or identify problems in an AI-generated answer. This matters in legal education because students need to reason independently and apply legal principles.
The mechanism is to ask students to show their thinking in more than one way. An oral explanation lets them describe why an answer follows from particular legal authorities. Finding errors in an AI answer asks them to examine the content critically rather than accept it at face value. Surya Kant specifically suggested these checks alongside written assessments.
He raised them while warning that AI detectors can wrongly flag students’ work. The article does not claim these methods guarantee a perfect assessment. They are proposed ways to supplement written work and avoid relying on a detector alone, while still checking students’ understanding and judgment.
What does disclosing AI assistance in legal research involve?
Disclosing AI assistance means making clear that AI was used while carrying out legal research. This would help readers, teachers or editors understand the role the tool played in the work. The article does not specify exactly what a disclosure should include, such as a required format or level of detail.
Surya Kant asked CLEA to develop model guidelines on disclosing AI assistance in legal research. The proposed guidelines would be available for adoption by law schools and journals across the Commonwealth. The article presents disclosure as part of setting clear rules for AI use, alongside teaching students the difference between using AI as an aid and letting it replace their own thinking.
The request is a proposal, not a report that guidelines already exist. If developed and adopted, such guidance could give institutions and journals a shared approach to disclosure. The article does not say when CLEA might produce guidelines or what specific requirements they would contain.
What is generative AI, and why can it produce plausible but incorrect legal answers?
The article does not define generative AI. In general, generative AI refers to systems that produce new content, such as written answers, in response to a prompt. This background definition comes from established general knowledge, not from the article. The article itself refers to AI-generated answers and warns that students should be able to identify errors in them.
Such systems can produce text that sounds confident and coherent without guaranteeing that every claim is accurate. In legal work, a plausible answer may still contain an error in its explanation or in how it applies a legal principle. The article does not explain the technical reasons for such errors; it stresses the need for students to think independently and check AI output.
Surya Kant proposed exercises in which students identify errors in AI-generated answers, as well as oral explanations and defences of legal authorities. PS Narasimha warned that AI’s greatest risk is replacing human thinking rather than assisting it. Their remarks frame AI output as something to assess, not accept uncritically.
This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.
Read more in the JupiteX app
Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.
Or read more news on the web