引導式測驗 · 中等 等級
人工智慧仲裁測驗
測試負責任的人工智慧在仲裁中的使用:允許的協助、保密性、機構規則、記錄驗證、揭露和仲裁員
問題 1 的 8
人工智慧工具總結了雙方提交的內容。誰負責決定爭議和最終裁決?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- 人工智慧工具總結了雙方提交的內容。誰負責決定爭議和最終裁決?
- The tool provider that generated the summary
- The party whose brief was summarized first
- The arbitrator who evaluates the record and issues the award
- The institution’s software administrator
- 在將展品上傳到雲端人工智慧服務之前,哪項檢查最重要?
- Review the agreement, procedural orders, rules, and provider data terms for confidentiality requirements
- Assume all arbitration material is public because hearings are private
- Upload first and ask the institution to approve later
- Remove the case caption but retain all sensitive details
- 產生的案例引文出現在裁決草案中。需要什麼驗證?
- Accept it if the citation format looks correct
- Ask the same model to confirm its earlier answer
- Keep it if both parties cited similar cases
- Check the cited authority and proposition against reliable legal sources before relying on it
- AAA 將其人工智慧仲裁器描述為在兩方流程中產生建議的裁決。哪個差異重要?
- The system’s output automatically binds every arbitration
- The description includes a human arbitrator issuing the award
- The tool replaces the parties’ agreement
- The product description sets rules for all institutions
- 摘要忽略了一方最強烈的相反論點。哪種品質檢查最能揭露這一點?
- Check only whether the summary is shorter than the briefs
- Measure grammar and readability alone
- Compare the summary with both submissions and test whether material positions and qualifications were omitted
- Ask the model to make the summary more confident
- A party wants an opportunity to address an AI-generated analysis that may affect the outcome. Which principle is implicated?
- The arbitrator’s ability to search documents privately
- The provider’s preferred prompt format
- The system’s benchmark accuracy
- Procedural fairness and a meaningful chance to address material issues
- Why should generated summaries retain exhibit identifiers and page references?
- So reviewers can trace each claim back to the original evidence
- So the model can replace the hearing record
- So the summary becomes self-authenticating
- So parties do not need access to source exhibits
- Which task is generally easier to bound with a narrow workflow and human correction path?
- Autonomous resolution of contested credibility
- Automated scheduling support with a way to correct errors
- Final award issuance without arbitrator review
- Secret ranking of parties by likely outcome