為何律師冒著制裁風險也要使用虛假AI引用

in a televised new york appellate division hearing, justices publicly humiliated a lawyer who submitted a brief containing citations to non-existent cases and misrepresented legal principles. attorney michael sanders, representing a plaintiff in a sidewalk injury case, could not explain the source of the fake citations when questioned, while justice lasalle noted the court had seen identical issues the week prior. the dressing-down extended to opposing counsel for not flagging the errors, highlighting that courts are actively identifying and rejecting the 'epidemic' of lawyers using generative ai tools without proper verification. this incident confirms that the legal system is moving from polite warnings to active scrutiny, with real-time sanctions becoming a risk for practitioners relying blindly on llm outputs.