Ohio Legal Community Weighs AI's Promise Against Real Risks
More than 90% of attorneys worldwide now use at least one AI tool in daily work, according to a March survey. Yet the legal profession remains cautious about where and how these systems belong in practice.
At an Ohio State Bar Foundation event in Columbus, judges, attorneys, and law professors discussed AI's growing role in legal work. The conversation revealed a clear divide: AI excels at specific tasks but should never replace human judgment in decisions that affect people's lives.
Where AI Works in Legal Practice
Legal research stands out as the most straightforward application. AI tools built into research platforms check for mistakes, find inconsistencies in documents, and quickly locate relevant sections of case records. Judge Scott Schlegel of the Fifth Circuit Court of Appeal in Louisiana, who has tested AI systems and speaks nationally on the topic, said these tools genuinely help attorneys and courts.
Fast courtroom transcripts represent another practical benefit. Faster transcripts allow attorneys to draft motions sooner, which speeds cases through the justice system and reduces court backlogs.
In chambers, judges and law clerks can use AI to organize case materials. Schlegel's appellate court uses AI to "chunk the record" - breaking it into sections by topic - and to compile witness lists and evidence summaries. These summaries save time during opinion writing.
But there's a catch. Judges and their staff must read through the original materials to verify AI summaries are accurate. AI can misunderstand context in testimony, Schlegel said, which he called "the most dangerous hallucination that's not talked about" in law.
Where AI Doesn't Belong
Judges should never use AI to draft initial opinions. Schlegel explained the risk plainly: if AI starts in the wrong direction, it becomes difficult to recover. "The tail wags the dog," he said.
Judicial decisions require human judgment. Schlegel pointed to a hypothetical: one AI system grants custody to one parent in a custody case, while another grants it to the other. Courts cannot outsource these decisions to competing algorithms.
"It's not our job to be first, it's our job to get it right," Schlegel said. "When a judge makes a mistake, that's the law, that's precedent."
The Hallucination Problem
Over 350 documented cases in the United States involve attorneys who filed legal documents containing fabricated citations. AI systems sometimes generate false information with confidence - a phenomenon called "hallucination."
Jeremy Kahn, AI editor for Fortune magazine, discussed these cases at the Columbus event. The problem underscores why verification matters in every AI application, no matter how routine the task appears.
Law Schools Face a Teaching Problem
Rebecca Fordon, a professor at Moritz College of Law who teaches legal research and technology, said law schools confront a genuine dilemma. As students use AI and see decent results, they trust it more. That trust can become dangerous.
Students must learn legal research skills before they learn to use AI. Fordon said instructors now assign in-class work specifically to prevent students from relying on AI tools prematurely.
"Students must develop judgment and analysis," Fordon said. "It's not enough to click 'verify' in a legal research tool. Students have got to read the cases and make the legal analysis complete."
Schlegel reinforced this point. Attorneys and judges who forget why they practice law - to protect people's lives, children, and livelihoods - will struggle most when using AI and risk harming the profession.
Broader Questions Remain
The Columbus panels also addressed ethics compliance when using AI, evidence verification in an age of AI-generated content, accessibility for people with disabilities, automation of tasks like deposition transcription, the risk of unauthorized practice of law, and whether AI could lower legal service costs.
The Ohio State Bar Foundation recorded the full discussion and made it available to the legal community.
For attorneys looking to understand how AI fits into legal practice, AI for Legal covers practical applications in research, document review, and compliance. Those in paralegal roles may benefit from an AI Learning Path for Paralegals that covers document automation and contract analysis.
Your membership also unlocks:
