New Delhi: The Delhi High Court Thursday allowed a plea by a homebuyers’ association to be withdrawn after senior advocates Sanjoy Ghose and N. Hariharan flagged the use of generative artificial intelligence chatbot ChatGPT in the petition.
A single-judge bench of Justice Girish Kathpalia was hearing the petition by a group of homebuyers challenging three orders of the Karkardooma court from earlier this month on the grounds that they were arbitrary, passed hastily and procedurally irregular in nature.
During the course of the arguments, however, it struck the senior lawyers, representing the respondents, that the petitioners in this case, representing Greenopolis Welfare Association, which claimed to be a homebuyers’ association, had erroneously referred to fake quotes and cases.
The lawyers pointed out to the bench that several judgments in the respondent’s plea were cited from ChatGPT, and asked the court for permission to initiate action against the erring lawyer.
The respondent’s lawyers also submitted before the court an entire table of cases cited by the petitioners, which the court recorded in its order. On one side, the respondent homebuyers put a column of what the petitioners had pleaded or argued, and on the other side, a column detailing the correct or actual position of law.
The high court dismissed the plea as withdrawn after orally observing that the use of AI in such a manner was “impermissible”, and a cause of serious concern.
Senior advocate Ghose told ThePrint, “We are shocked that someone had used ChatGPT in this manner. It wasn’t just a typographical error, but they had manufactured cases, and used ChatGPT’s inaccurate information, without any verification or cross-checking, as should have been done.”
Calling this a very serious issue, the court said that it would proceed in accordance with the law, which, according to Ghose, means it would either take contempt action against them, or prosecute them for perjury.
How ChatGPT usage was detected
Advocate Ajay Sabharwal, who also appeared for the respondents, told ThePrint that they, in fact, pointed it out to the court that all paragraphs were either false or fabricated. “The case laws cited by them were entirely non-existent, and we submitted that this amounted to contempt of court,” he said, adding that the court took note of these submissions and remarked that it was a serious matter.
For instance, one of the judgments cited by the Gurugram-based Greenopolis Welfare Association was the landmark Supreme Court ruling in Raj Narain vs Indira Gandhi case (1972).
The respondents pointed out that although the petitioners had cited paragraph 73 of the 1972 ruling, the judgment had only 27 paragraphs, so there was no question of there being a 73rd paragraph in the picture. Such a paragraph as was quoted in the petition was entirely fabricated and did not exist, the petitioners told the court.
Similarly, the petitioners had cited paragraph 21 of the 2010 ruling in Dalip Singh vs State of UP, which, according to them, read, “In the last 40 years, a new creed of litigants has cropped up. Those who belong to this creed do not have any respect for truth. They shamelessly resort to falsehood and unethical means for achieving their goals.”
However, the respondents pointed out that the actual para 21 of the 2010 ruling read differently: “A perusal of the application dated 8-7-1976 submitted by Shri Praveen Singh.” Notably, the para cited as para 21 was actually para 2 of that case.
Also Read: Did SC judge condone delay? Legal tug of war over control of Delhi Gymkhana takes peculiar turn
Not the first time
However, this is not the first time that a court has flagged the use of AI by lawyers in arguing cases. While some courts and judges have come down heavily on the use of AI in judicial work, others, like Manipur High Court and Punjab and Haryana High Court, have leveraged its use for research purposes, while acknowledging the same in their orders.
Manipur High Court had turned to ChatGPT for research in a service law matter last year in May in a case concerning reinstatement of Village Defence Force (VDF) personnel. But Chief Justice of India B.R. Gavai, in March this year, flagged the inherent risks posed by the use of AI in such cases.
“Relying on AI for legal research comes with significant risks, as there have been instances where platforms like ChatGPT have generated fake case citations and fabricated legal facts,” the CJI had said, while speaking at the Kenyan Supreme Court in Nairobi.
In July this year, the Kerala High Court came out with guidelines for the district judiciary concerning the use of generative AI for judicial work.
“Submitting information such as facts of the case, personal identifiers or privileged communications or uploading any other documents relating to the litigations to any such Al tools may result in serious violations of confidentiality. Hence, the use of all cloud-based services should be avoided, except for the approved AI tools,” the official guidelines read.
In 2023, Justice Pratibha Singh, while acting on Christian Louboutin’s case, had refused to rely on ChatGPT’s response presented by the luxury shoe brand. “The above responses from ChatGPT as also the one relied upon by the Plaintiffs shows that the said tool cannot be the basis of adjudication of legal or factual issues in a court of law,” the court had said.
The response of Large Language Model-based chatbots, like ChatGPT, comes with possibilities of incorrect responses, fictional case laws, imaginative data etc. generated by AI chatbots, the court had also noted.
“Accuracy and reliability of AI generated data is still in the grey area. There is no doubt in the mind of the Court that, at the present stage of technological development, AI cannot substitute either the human intelligence or the humane element in the adjudicatory process. At best the tool could be utilised for a preliminary understanding or for preliminary research and nothing more,” the court had said.
This is an updated version of the report.
(Edited by Mannat Chugh)
Also Read: Veera Raja Veera row: Why Delhi High Court dismissed copyright case against AR Rahman