Q: What does the Punjab Haryana HC AI ban require judicial officers to do? A: The Punjab Haryana HC AI ban directs judicial officers to avoid using AI tools such as ChatGPT, Gemini, Microsoft Copilot and Meta AI for writing judgments or conducting legal research. The instruction is to be communicated through district and sessions judges and warns that any violation will be viewed seriously. Q: Which AI tools and features are covered by the order? A: The order names ChatGPT, Gemini, Microsoft Copilot and Meta AI and applies to any similar AI tool. It covers web chatbots, AI assistants in browsers or office software, and plugins that auto-suggest legal arguments or citations. Q: Who must comply with the ban and how should compliance be confirmed? A: All judicial officers working under the High Court’s control must comply, with district and sessions judges responsible for issuing the instruction. Officers are expected to confirm in writing that they have received the direction and have disabled AI features where applicable. Q: Why are courts cautious about using AI in judicial decision-making? A: Courts are cautious because AI can invent cases or citations, operate opaquely, reproduce bias, risk leaking confidential case material to external servers, and shape precedent in ways lower courts may follow without scrutiny. Justice Ashwani Kumar Mishra and other judges have warned that premature adoption could create systemic risks and a crisis in judicial dispensation. Q: What practical steps can judicial officers take now to comply while keeping workflows efficient? A: Officers can disable AI assistants in word processors, browsers and email clients, remove drafting extensions, and block AI domains on chamber networks to avoid accidental uploads. They should rely on verified legal sources, verify every citation with primary material, protect sensitive files in secure court systems, and keep a simple research log noting that no AI tools were used. Q: Which digital tasks may still be acceptable under the ban? A: The direction bars AI for judgments and legal research but allows non-AI digital tools for administrative tasks such as scheduling, file tracking, document formatting, accessing official cause lists and eCourts services. When a feature is labeled “AI” or it generates content or performs analysis, officers should avoid it and seek guidance before adopting the tool. Q: How does the ban affect lawyers and litigants who use AI for research or drafting? A: Counsel should not rely on AI-generated research or submit AI-produced citations and must provide clear sources and copies of authorities cited. Case data and pleadings should not be shared with public AI tools, and courts may scrutinize submissions more closely for accuracy. Q: What longer-term safeguards are courts considering before adopting AI in adjudication? A: Longer-term safeguards under consideration include clear rules defining where AI is never allowed and where limited use is permitted, vetted secure tools running on controlled infrastructure with audit logs, and bias testing with independent reviews. Training for judges and staff, red-teaming, and a strict human-in-the-loop requirement are also part of the proposed guardrails before broader adoption.