Singapore lawyers sanctioned for AI-generated false cases; OpenAI sued over ChatGPT legal advice
Singapore courts imposed sanctions on two lawyers who submitted AI-generated fabricated court cases in legal filings, while in the US, Japanese life insurance company Nippon Life sued OpenAI for ChatGPT providing erroneous legal advice that caused the company significant losses. Both incidents highlight risks from unverified AI-generated legal content and raise questions about AI use in regulated legal services.
What happened
In Singapore, two lawyers—Wu Bishan with over 30 years of practice and Amarjit Singh Sidhu—submitted AI-generated fabricated court cases in closing arguments in a civil lawsuit over assets of deceased businessman Chan Chuan De. When the plaintiff discovered two fictional cases in the defence's closing arguments, the lawyers claimed a paralegal subordinate had used AI tools without their knowledge, and that the employee had since left and was uncontactable.
In the US, Nippon Life Insurance sued OpenAI on March 4 over ChatGPT providing improper legal advice that a customer used to draft legal documents containing serious errors, causing the company to incur substantial legal costs defending against resulting disputes.
Reported impact
The Singapore High Court found that the fabricated cases wasted the plaintiff's legal resources, consumed court time, and created risk of unjust judicial outcomes. The court emphasized the cases posed a serious threat to Singapore's judicial system.
Nippon Life Insurance claims ChatGPT's defective legal advice caused the company to expend significant resources in litigation to address the resulting legal disputes.
Index analysis
In Singapore, the court determined both lawyers failed their professional duty to verify all materials submitted to court. The judge noted lawyers bear responsibility for ensuring accuracy regardless of whether AI tools assisted in preparation, and found insufficient supervision mechanisms over subordinates' use of AI.
The US lawsuit argues ChatGPT provided legal services equivalent to unlicensed legal practice, as only licensed attorneys may provide legal advice under American law. OpenAI's defence states ChatGPT is a tool and users must verify outputs; the company's terms prohibit using ChatGPT for legal advice.
Governance controls
Singapore's courts have issued guidance on AI use since October 2024 requiring all parties, lawyers and witnesses to ensure AI-generated content is accurate and truthful, prohibit fabrication of false cases, and subject violators to costs orders and potential further legal action. The High Court judge noted courts cannot oppose AI tool use but must enforce that technology plays only an assisting role with lawyer verification.
OpenAI's terms of service explicitly prohibit users from employing ChatGPT for legal advice. Legal experts predict the lawsuit could establish an AI legal liability precedent that drives stricter use policies across regulated sectors including law, healthcare and finance.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.