ChatGPT in Court: Risks and Rewards for Litigants in Person

The New Counsel in Town: AI in the Courtroom

A quiet but profound transformation is underway within the hallowed halls of British and Irish justice. It is not being driven by legislative reform or a new generation of barristers, but by a force accessible to anyone with an internet connection: artificial intelligence. The long-standing figure of the Litigant in Person (LiP)—an individual navigating the complex legal system without a solicitor—is evolving. We are now witnessing the emergence of the ‘AI-augmented litigant’, a self-represented party armed with powerful Large Language Models (LLMs) that act as their tireless, if sometimes unreliable, digital paralegal.

For decades, the LiP has been a feature of our courts, often driven by necessity rather than choice. However, the tools at their disposal have changed dramatically. Where once they might have relied on paper templates from Citizens Advice or the goodwill of a ‘McKenzie Friend’, today they are turning to platforms like OpenAI’s ChatGPT, Google’s Gemini, and Anthropic’s Claude. These generative AI systems are reshaping the landscape of self-representation, democratising access to legal information while simultaneously introducing novel and significant risks that the judiciary is only now beginning to confront.

The Perfect Storm: Why More People Are Representing Themselves

The surge in LiPs is not a sudden phenomenon but the result of converging societal pressures. The primary catalyst remains the prohibitive cost of legal services. A protracted High Court dispute can easily generate fees running into the tens of thousands of pounds, a sum far beyond the reach of the average household. This has created a vast ‘missing middle’—a demographic of individuals who earn too much to qualify for the stringent means-testing of civil legal aid, yet not nearly enough to afford private representation without risking financial ruin.

Compounding this is the evolution of public attitudes towards formal institutions. The modern claimant is more technologically literate and less intimidated by the prospect of self-advocacy than their predecessors. The ‘Google Law’ era, where individuals would search for legal answers online, has seamlessly transitioned into the ‘AI Law’ era. The perceived accessibility of these advanced tools gives many the confidence—sometimes misplaced—to believe they can level the playing field against opponents with deep pockets and extensive legal teams. The result is an unprecedented number of people choosing to go it alone, with an AI as their shadow counsel.

From Frustration to Formal Pleadings: AI’s Role in Litigation

For the uninitiated, the procedural and linguistic conventions of the legal world are a formidable barrier. This is where generative AI has proven to be a game-changer for LiPs. Its applications are as varied as they are powerful. An individual can input a rambling, emotional narrative of their grievance, and within seconds, the AI can restructure it into a coherent Statement of Claim, formatted correctly and employing the appropriate legal terminology.

The technology excels at demystifying complexity. An LiP faced with a 50-page judgment can ask an LLM to summarise its key findings and legal principles into a few digestible bullet points. The labyrinthine Rules of the Superior Courts, which can baffle even seasoned legal professionals, become navigable with an AI that can explain, on demand, the purpose of a Notice for Particulars or the correct procedure for filing an Affidavit of Service. Furthermore, AI acts as a crucial emotional filter. It can transform a frustrated, angry email into a detached, formal ‘without prejudice’ letter, helping the LiP maintain the professional tone required for effective legal correspondence. In essence, AI provides the scaffolding of legal competence, allowing laypeople to present their cases with a veneer of professionalism previously unattainable.

The Phantom Precedent: When AI ‘Hallucinates’ Justice

This newfound power, however, comes with a perilous flaw. LLMs are not repositories of legal truth; they are incredibly sophisticated statistical engines designed to predict the most plausible sequence of words. This can lead to a phenomenon known as ‘hallucination’, where the AI confidently generates entirely fictitious information. In a legal context, this is catastrophic. The judiciary in both the UK and Ireland has raised the alarm over a growing number of cases where submissions have been built upon a foundation of non-existent case law.

The cautionary tale that reverberated through the legal profession was the UK case of Harber v HMRC in 2023. A litigant in person, seeking to support her tax appeal, presented the tribunal with citations for nine previous cases. They appeared impeccable, complete with names, dates, and concise summaries of the supposed judgments. Yet, upon investigation, it was discovered that not a single one of them was real. The AI, when prompted to find supportive case law, had simply invented it. More recently, the Judicial Council of Ireland issued formal guidelines in late 2025 after a series of incidents, including a Workplace Relations Commission hearing where a claimant’s submissions were found to be riddled with phantom citations. This poses a direct threat to the integrity of the judicial process, wasting court time and undermining the very basis of a common law system built on precedent.

The Judicial Response: Pragmatism Over Prohibition

The courts are now grappling with how to manage this technological disruption. Outright prohibition is seen as both impractical and potentially detrimental to access to justice. A notable moment of judicial pragmatism came in mid-2025 when an Irish High Court judge, Mr Justice Barrett, declined to criticise a party for using AI-generated correspondence. He observed that there is no legal principle requiring a solicitor’s letter to be answered by a human-authored one, acknowledging the inevitable integration of such tools.

However, the emerging consensus for 2026 is one of strict accountability. The judiciary has made it unequivocally clear that the litigant bears absolute responsibility for the accuracy of every word submitted to the court. The excuse that “the AI made it up” holds no water before the Master of the High Court or a circuit judge. Verification is paramount. The convenience of AI does not absolve the user of the fundamental duty to ensure their submissions are factually and legally sound. The onus is now firmly on the AI-augmented litigant to become a diligent fact-checker.

The Future: A New Era of Legal Literacy

We are at a critical juncture where technology and tradition are colliding in the courtroom. AI has the undeniable potential to bridge the justice gap, empowering ordinary people to articulate their rights more effectively. Yet, this potential can only be realised if it is paired with a new form of digital legal literacy. The modern Litigant in Person must learn to use these tools defensively, treating the AI not as an oracle but as a talented but untrustworthy assistant.

This means cultivating a habit of rigorous cross-referencing, checking every case citation against official databases like BAILII or the Irish Statute Book. It means using AI for structure and tone but supplying the core facts and evidence oneself. Above all, it requires understanding that AI is a preparatory tool, not a strategic one. It cannot replicate the nuanced judgment, ethical considerations, or the intuitive ‘gut feeling’ of an experienced human lawyer. The courtroom of tomorrow will undoubtedly feature more technology, but for the litigant navigating it alone, success will depend on remembering that in a court of law, the only currency that matters is not what is plausible or probable, but what is true.