b'THINKING ABOUT USING GENERATIVE AI? JUDICIAL AI UTILIZATION GUIDELINESEthical Oversight - The Code of Attribution -Ensure your workJudicial Conduct applies fully to the product does not infringe uponuse of AI technologies. You maintain copyright or intellectual property rights ultimate responsibility for all rulings by proper attribution to sources asand legal documents.necessary.Competence - The judicial branch Confidentiality - If using a publicmust stay informed about evolving AI generative AI tool (like ChatGPT), yourtechnologies. Prior to using any input prompt is being handed over totechnology, including generative AI the technology. Ensure you do notapplications, you need to understand compromise sensitive information. Doboth general AI capabilities and the not input any information such asspecific tools being used.(non-exhaustive list):Confidential or privilegedAccuracy - Be mindful that content information;from generative AI applications comes Personal identifying information;from sourced material. AI-generated Protected health information;content must be thoroughly reviewed Justice and public safetyto ensure accuracy and compliance information;with legal and ethical obligations, Code containing passwords orincluding a specific need to guard security-related information; andagainst technology lending to Information that has potential tounintentional bias or prejudice. erode public trust. WHAT TO WATCH FOR AS A JUDGEAI is ubiquitously present in modern technology. It is safe to assume AI was used in pleadings andother written materials. Generative AI applications gives rise to these considerations:Hallucinations - When generative AI produces output that appears CHECK CITATIONSrealistic but is misleading or made up by the AI itself rather than real-world data or input. If generative AI was used in a legal pleading orbrief, included citations may be entirely made up or be a real case butnot contain the purported language cited. Read a case decision onhallucinations. HEAR EVIDENCE ONDeepfakes - When generative AI is intentionally used to produce FOUNDATIONSconvincing/deceptive media - images, audio, video, etc. Fake evidenceis relatively easy to create and reliable technology solutions do notexist to identify real vs fake evidence. Learn more about identificationof deepfakes.ENFORCE EXISTINGExtended Reality - Technology can seamlessly be integrated with our TECHNOLOGYperson though wearable devices. Recognize that technology may RULESallow an individual to record and analyze audio and video and channelinformation from other sources in real-time, potentially withoutdetection. Learn more about extended reality.'