When Artificial Intelligence Walks Into the Courtroom: Mariam Shatberashvili on Artificial Intelligence, Access to Justice, and the Future of Legal Systems

Mariam Shatberashvili, J.D., LL.M., MBA

Artificial intelligence is no longer a futuristic concept within the legal profession. In many respects, artificial intelligence has already walked into the courtroom. It is performing functions traditionally associated with attorneys: drafting motions, summarizing records, assisting with legal research, reviewing contracts, and influencing litigation strategy. Yet as courts increasingly confront fabricated citations, hallucinated case law, and broader concerns surrounding accountability, an important question continues to emerge: who bears responsibility when artificial intelligence contributes to legal decision-making?

In many respects, artificial intelligence is addressing a problem the legal profession has struggled with for decades: access to legal information. Individuals who cannot afford attorneys charging hundreds of dollars per hour are now able to ask AI platforms legal questions, generate pleadings, organize arguments, and navigate systems that have historically been financially and procedurally inaccessible. For many self-represented litigants, particularly in housing disputes, debt collection matters, immigration proceedings, and family law cases, artificial intelligence may represent the first meaningful form of legal assistance available to them. The significance of that development should not be understated.

At the same time, however, these technologies introduce an entirely new institutional concern.

Artificial intelligence systems are not reliable in the manner legal institutions require legal authority to be reliable. AI platforms are capable of fabricating cases, inventing quotations, misstating legal standards, and generating inaccurate legal analysis while presenting that information with apparent confidence and professionalism. The concern becomes particularly significant when AI-generated filings appear sufficiently polished that courts may not immediately recognize that the authorities cited do not actually exist. The issue gained national attention following Mata v. Avianca, Inc., where attorneys submitted fictitious legal authorities generated through artificial intelligence tools.

This creates a significant tension between accessibility and institutional reliability.

On one side, society is moving toward broader access to legal guidance through technology. On the other side, courts, judges, clerks, and attorneys are increasingly forced to verify whether the legal authorities being cited actually exist. The burden on the judicial system no longer involves solely evaluating legal arguments or procedural sufficiency. It now also includes verifying the authenticity and existence of the underlying authorities themselves.

Within already overburdened judicial systems, that challenge is substantial.

Rejecting artificial intelligence entirely is neither realistic nor practical. These technologies are already transforming how individuals interact with legal systems, and that transformation will likely continue to accelerate. At the same time, uncritical reliance on AI-generated legal work risks undermining the accuracy, reliability, and institutional trust upon which legal systems depend.

The central challenge moving forward will likely be achieving an appropriate balance between accessibility and reliability.

Legal institutions may ultimately need to develop hybrid frameworks in which artificial intelligence increases accessibility to legal information while human professionals remain responsible for oversight, verification, ethical judgment, and accountability. Artificial intelligence may increasingly serve as a preliminary legal assistance tool rather than a substitute for professional legal reasoning itself. Courts may eventually require disclosure standards for AI-assisted filings, stricter verification obligations, or expanded ethical frameworks governing the use of generative artificial intelligence in litigation. Organizations such as the American Bar Association and the National Center for State Courts have already begun examining how legal institutions may need to adapt to the increasing use of generative AI within judicial systems.

Artificial intelligence can provide efficiency, organization, and broader access to information. What it cannot independently provide is ethical judgment, accountability, or institutional responsibility. Those functions remain dependent upon human oversight.

And that may ultimately become the defining issue once artificial intelligence fully walks into the courtroom: not whether these systems can participate in legal processes, but whether legal institutions can integrate them without compromising reliability, procedural fairness, and public trust in the administration of justice.

Further Reading & Resources

• Legal Services Corporation – The Justice Gap Report

• American Bar Association – Artificial Intelligence Resources for Legal Professionals

• National Center for State Courts – AI and the Courts

• Stanford Human-Centered Artificial Intelligence (HAI)