For its part, the Hawaii Supreme Court thinks it has the resources already. Rule 11 of the Hawaii Rules of Civil Procedure, which mandates that statements made to the court be truthful and permits judges to impose sanctions for violations, is “broad enough to provide adequate safeguards” against AI misuse, according to a long-awaited report presented to acting Chief Justice Sabrina McKenna. It’s a bold assertion. Additionally, an increasing number of legal scholars find it difficult to accept.

The issue is not theoretical. Carlos Valencia, a lawyer, filed a brief with artificial intelligence-generated citations. At about the same time, Kaʻōnohiokalā J. Aukai IV, another attorney from his firm, submitted a document in which all six citations were faulty, with two of them being wholly made up and the others misrepresenting actual cases.

He expressed regret. It was approved by the judge. No penalties. In a different case, the presiding judge reprimanded Honolulu Deputy Corporation Counsel David Sgan for using AI-generated fake law in a civil rights case against the Honolulu Police Department. Just a reprimand.

As these cases mount, it seems like the repercussions haven’t kept up with the behavior. On paper, Rule 11 is in place. Its application has been, at best, uneven, at least in state courts in Hawaii. At the time of the report’s publication, French researcher Damien Charlotin’s database contained 690 recorded instances of this type of failure worldwide, including 470 in the US. Since late August, that figure had more than doubled. Slowing down is not the issue.

The federal courts in Hawaii have adopted a very different strategy. When AI is used to create a court document, attorneys must disclose it and make sure the content isn’t fake. There are actual penalties for violations. It is hard to overlook the difference with state court rulings, and the Hawaii Supreme Court’s report states that it is currently contemplating a similar disclosure requirement. However, this consideration is not yet a rule.

The report itself is measured, at times even thoughtful. It recognizes the real potential of AI tools, including their ability to assist non-native English speakers in navigating the legal system, process documents, produce transcripts more quickly, and increase access for those who cannot afford attorneys. It would be unjust to ignore these genuine advantages.

However, the same report also states unequivocally that “if attorneys fail to exercise independent judgment and fail to validate and verify outputs from AI tools,” there will be chaos, wasted resources, unsuccessful advocacy, and false claims made in court. The conclusion that current regulations are adequate and that warning go hand in hand.

In contrast, Illinois provided an example of what accountability might entail. Lawyers filed a post-trial motion full of AI-generated errors in a case involving the Chicago Housing Authority, where a jury had already returned a $24 million verdict after concluding that children’s exposure to lead in the authority’s housing caused permanent brain damage.

Fourteen citations were found to be incorrect by opposing counsel. False citations were deemed “a grave threat to the judicial branch” by the court, which imposed $59,500 in fines. The decision stated that decisions made by courts must be based on real law, not fiction, in order for the public to have faith in them. It’s difficult to ignore the fact that Hawaii still lacks that degree of clarity.

For his part, Valencia informed the court that any brief that associates submit for his consideration must include copies of the cited cases. “This kind of mistake will not happen again,” he said. That might apply to his practice. It is far less evident that this is the case for the profession or for a court system that is still debating the adequacy of its current regulations.

The post Hawaii’s Supreme Court Says Existing Rules Already Cover AI Abuse. Legal Scholars Disagree appeared first on Creative Learning Guild.