In the rapidly evolving world of legal technology, a peculiar issue has arisen: lawyers are increasingly citing non-existent cases, thanks to errors generated by AI tools. While the primary responsibility for these "hallucinated" citations lies with attorneys who fail to verify their sources, the problem is exacerbated by the proprietary citation formats used by major legal databases like Westlaw and LexisNexis.

During the recent Legaltech Week roundtable, attention was drawn to a blog post by law librarian Shay Elbaum, discussing a notable instance in the Flycatcher Corp. v. Affable Avenue case. Here, a lawyer argued unsuccessfully that AI-generated proprietary citations are difficult to verify without access to the specific database. Despite having access to Westlaw, the lawyer's argument highlighted a broader issue affecting those without such resources.

The reliance on proprietary formats not only impacts the quality of legal arguments but also places undue vulnerability on opposing counsel, particularly those who cannot afford expensive database subscriptions. This issue was starkly illustrated by a review of Damien Charlotin’s database of recent legal filings, which revealed numerous instances of fabricated "WL citations" supposedly referencing unpublished cases accessible only through Westlaw.

The situation calls for a robust response from both publishers and the legal system. Proposed solutions include requiring publishers like Thomson Reuters and LexisNexis to allow verification of citations through a free public lookup, even if it means stripping out some proprietary features. Moreover, an adaptation of Federal Rule of Appellate Procedure 32.1(b), which mandates attorneys to provide copies of cited unpublished cases, could be expanded to other jurisdictions and court rules to level the playing field.

Despite these suggestions, some districts that have implemented similar rules still experience non-compliance, indicating a need for stricter enforcement and perhaps a more comprehensive approach to tackle the issue.

As AI continues to integrate into legal research tools, the industry must address these "hallucinations" not just as individual failings but as a systemic risk that could potentially undermine judicial integrity if left unchecked. The legal community must take proactive steps to ensure that what appears on paper is as real and verifiable as the law itself.