Former lord chancellor Sir Robert Buckland KC urged courts to employ the costs regime to curb the flood of “AI slop” produced by litigants in person, warning that unchecked use threatens the justice system.
AI-generated material strains the courts
Speaking at the Costs Law Reports conference, Buckland said AI has created a profound shift in litigation economics, moving expenses from the creator to opponents, courts and taxpayers. He noted that the speed of generation reduces the traditional time-consuming drafting process, which previously required hours of careful work.
He explained that a self-representing party can feed a grievance to an AI tool and receive a pleading, witness statements, a skeleton argument, and a list of authorities, often in minutes. The resulting documents often contain extensive citations, but many of those references are either irrelevant or duplicated, which may increase the burden on the court to filter them.
According to Buckland, the person generating the material incurs virtually no marginal cost, while the adversary, the court and the public bear the resulting expense. He pointed out that this shift places a hidden cost on public funds, as court resources must be allocated to review and manage the surplus material. He noted that before AI, prolixity was naturally limited because a lawyer’s fee acted as a check.
Read Also: Clinical Negligence Claims Against NHS Reach Record Settlement Times
Buckland argued that the existing powers under CPR 44 allow judges to impose costs orders for unreasonable conduct, and AI-generated conduct falls within that definition. The rule already permits sanctions for conduct that unnecessarily inflates litigation costs, and AI output fits that description when used without proper oversight.
He cautioned that a blanket approach of ‘LIP + AI = costs sanction’ would be unfair and unworkable. He argued that courts should exercise discretion, including issuing warnings before imposing penalties, to ensure fairness for individuals accessing technology.
Distinguishing uninformed from irresponsible AI reliance
He added that a distinction must be drawn between uninformed and irresponsible AI use. A person who sincerely believes an AI proposition is correct may need education. In contrast, a person told an authority does not exist yet continues to rely on it behaves very differently.
He noted that a litigant who submits a long document because they genuinely cannot understand the court’s requirements is not in the same position as one who deliberately floods the court with irrelevant material after having been warned to stop.
Read Also: Lawyers vow to continue fight for commission transparency
Proposed procedural safeguards
Buckland suggested that unreasonable costs generated by the use of AI may be taken into account in determining the amount of costs, providing a principled basis for intervention. Such consideration would align with existing case law that ties cost awards to the behaviour of the parties throughout the proceedings.
Professional lawyers cannot claim to be unaware of the dangers of relying on AI. Presenting a document generated by ChatGPT, Claude, Gemini, or another system does not excuse responsibility. He warned that reliance on a single AI system without verification could lead to inadvertent errors that compromise the integrity of the case.
Career background and current roles
Buckland served as lord chancellor from 2019 to 2021 under Prime Minister Boris Johnson. He now practices as a barrister at Foundry Chambers and acts as a consultant for DAC Beachcroft.
