@techreport{oai:ipsj.ixsq.nii.ac.jp:00234328,
 author = {中川, 裕志 and Hiroshi, Nakagawa},
 issue = {2},
 month = {May},
 note = {代理人が行う行為のうち,AI に可能なものと不可能なものを切り分けおくことは,Personal AI (PAI と略記) に本人の代理人が行う行為のどの部分を任せられるかの判断あるいは PAI の設計指針を検討するにあたって役立つ知見となる.ここでは,主にアメリカの代理法の第 3 次リステイトメントを参考にして,代理人に行う種々の行為が,本人の指示通りに動くツール PAI と本人の指示に従いつつも自律的に動く自律的 PAI で行えるかどうかという観点から分析を行う.本稿は,この論点について報告した「Personal AI が行う代理的行為についての考察その 1」の続報である., Separating out those acts that can be performed by an agent from those that cannot be performed by an AI will provide useful information in determining which parts of the acts performed by the principal's agent can be entrusted to a Personal AI (PAI) or in considering design guidelines for a PAI. In this paper, we mainly refer to the Restatement Third. Agency of the U.S. and analyze the PAI from the viewpoint of whether various acts performed by the agent can be performed by a tool PAI that follows the instructions of the principal or an autonomous PAI that follows the instructions of the principal but operates autonomously. This paper is a continuation of " Consideration of agency-like behavior by Personal AI Part 1" which reported on this issue.},
 title = {Personal AIが行う代理的行為についての考察その2},
 year = {2024}
}