06/10/2025
Courts worldwide are navigating uncharted waters with artificial intelligence, and their radically different approaches reveal a governance crisis that demands immediate attention from senior leaders. Across eight major jurisdictions, courts have responded to generative AI with starkly contrasting frameworks: New South Wales has imposed categorical prohibitions on AI-generated witness evidence and mandates sworn declarations that AI was not used,¹ whilst Singapore takes a permissive stance requiring no disclosure unless specifically requested, placing full responsibility on individual practitioners.² This fragmentation is not merely academic. Courts in the United States and Australia have already sanctioned lawyers for filing submissions citing entirely fabricated cases generated by AI 'hallucinations', where systems like ChatGPT created plausible-sounding but completely fictitious legal precedents.³ The consequences extend far beyond professional embarrassment to fundamental questions about evidentiary integrity, access to justice for self-represented litigants, and the preservation of confidential information that may be inadvertently fed into public AI systems and become permanently embedded in their training data.⁴The window for proactive governance is closing rapidly, yet no international consensus has emerged on how to balance innovation with risk management in the administration of justice. New Zealand has pioneered a three-tiered approach with separate guidelines for judges, lawyers and non-lawyers, recognising that different court users face fundamentally different obligations and capabilities,⁵ whilst the United Kingdom has focused exclusively on guidance for judicial officers without addressing practitioner conduct.⁶ For government executives responsible for policy development, regulatory frameworks, and public sector digitalisation, understanding these divergent approaches is not optional. The report exposes critical gaps in current governance models and demonstrates why courts are moving from permissive to restrictive regulation as verification mechanisms struggle to keep pace with technological advancement.⁷ Download the full analysis to understand how these judicial responses should inform your organisation's approach to AI governance, professional liability frameworks, and access to justice initiatives before fragmented regulation creates compliance nightmares across jurisdictions.
