AI’s Growing Role and Risks in the Modern Legal Landscape
Recent research covering 4.5 million federal civil cases from 2005 to 2026 shows a sharp rise in self‑represented filings: the share of pro se lawsuits grew from 11 % in 2022 to 16.8 % in 2025, and the absolute number of AI‑assisted submissions more than doubled after 2023. Judges like Magistrate Maritza Braswell in Colorado report that many of these documents bear the unmistakable hallmarks of large‑language‑model output – formulaic prose, occasional “hallucinated” citations, and even fabricated quotes. While AI tools can lower the cost of drafting pleadings and help litigants navigate complex procedures, the data suggest they do not improve win rates. Courts are now grappling with fundamental questions: should a chatbot owe a duty of care comparable to a human attorney, and who bears liability when an AI‑generated legal opinion proves erroneous?
A cautionary example emerged in the English Chancery Division case Cork and another v Smith [2026] EWHC 1199, where a junior associate relied on an AI system to locate the relevant insolvency rule. The AI invented a non‑existent provision – “Insolvency Rule 12.37(5)” – and even flagged the answer as potentially inaccurate, a warning that was ignored. The false text was submitted to the court, prompting the judge to intervene because of “misleading statements” from the solicitor’s earlier correspondence. This incident underscores a persistent problem: AI can produce plausible‑looking but false legal material, and without rigorous verification against primary sources, such hallucinations can jeopardize procedural fairness and increase the workload of already stretched judges.
Beyond individual cases, AI is reshaping strategic battles between corporations and activists. In early 2026 the long‑running Donovan–Shell feud entered an “AI‑driven bot war,” with the activist feeding decades of leaked documents into multiple chatbots to generate narratives that paint Shell as the losing side. At the same time, the Indian Supreme Court released a draft framework that treats AI strictly as an assistive tool – banning its use for adjudication or sentencing while encouraging its deployment for administrative efficiency and access to justice. Together, these developments highlight a paradox: AI promises broader access to legal resources but simultaneously introduces new sources of error, misinformation, and accountability challenges. Policymakers, courts, and legal professionals must therefore craft clear standards for AI verification, define liability for erroneous advice, and ensure that the technology augments rather than undermines the integrity of the justice system.
