How to Explain and Justify Almost Any Decision: Potential Pitfalls for Accountability in AI Decision-Making
How to Explain and Justify Almost Any Decision: Potential Pitfalls for Accountability in AI Decision-Making
复制标题
如何解释和证明几乎所有决策的合理性：人工智能决策中问责制的潜在陷阱
DOI：
10.1145/3593013.3593972
复制
发表时间：
2023
期刊：
影响因子：
--
通讯作者：
Joachims, Thorsten
中科院分区：
文献类型：
--
作者：
Zhou, Joyce;Joachims, Thorsten
关于"获得解释的权利"的讨论日益相关，因为它对审计自动化决策系统以及对此类决策提出异议具有潜在的效用。然而，大多数现有的解释工作集中在协作环境中，设计师的动机是实施善意的解释，揭示决策系统的潜在弱点。这种动机在审计环境中可能不成立。因此，我们要问：有多少解释可以被恶意地用来捍卫决策系统？在本文中，我们演示了如何黑盒解释系统开发来捍卫一个黑盒决策系统可以操纵决策接受者或审计员接受故意歧视的决策模型。在一个个案的情况下，决策接受者无法分享他们的情况和解释，我们发现，大多数个人的决策接受者可以收到一个可验证的理由，即使决策系统是故意歧视。在一个系统范围内的情况下，每一个决定是共享的，我们发现，虽然理由经常相互矛盾，没有直观的阈值来确定这些矛盾是因为恶意的理由，或者因为这些理由的简单性要求与模型行为相冲突。最后，我们讨论如何系统范围的指标可能是更有用的比解释系统评估整体决策的公平性，而解释可能是有用的公平性审计之外。
Discussion of the “right to an explanation” has been increasingly relevant because of its potential utility for auditing automated decision systems, as well as for making objections to such decisions. However, most existing work on explanations focuses on collaborative environments, where designers are motivated to implement good-faith explanations that reveal potential weaknesses of a decision system. This motivation may not hold in an auditing environment. Thus, we ask: how much could explanations be used maliciously to defend a decision system? In this paper, we demonstrate how a black-box explanation system developed to defend a black-box decision system could manipulate decision recipients or auditors into accepting an intentionally discriminatory decision model. In a case-by-case scenario where decision recipients are unable to share their cases and explanations, we find that most individual decision recipients could receive a verifiable justification, even if the decision system is intentionally discriminatory. In a system-wide scenario where every decision is shared, we find that while justifications frequently contradict each other, there is no intuitive threshold to determine if these contradictions are because of malicious justifications or because of simplicity requirements of these justifications conflicting with model behavior. We end with discussion of how system-wide metrics may be more useful than explanation systems for evaluating overall decision fairness, while explanations could be useful outside of fairness auditing.
登录
查看更多内容
“太多了？”
影响因子：
1.9
作者：
Joel J Heidelbaugh
通讯作者：
Joel J Heidelbaugh
影响因子：
30.8
作者：
Ghassemi, Marzyeh;Oakden-Rayner, Luke;Beam, Andrew L.
通讯作者：
Beam, Andrew L.
为什么和为什么不解释可以提高上下文感知智能系统的可理解性
DOI：
10.1145/1518701.1519023
发表时间：
2009
期刊：
Proceedings of the SIGCHI Conference on Human Factors in Computing Systems
影响因子：
--
作者：
Brian Y. Lim;A. Dey;Daniel Avrahami
通讯作者：
Daniel Avrahami
DOI：
--
发表时间：
2019-01
期刊：
--
影响因子：
--
作者：
U. Aïvodji;Hiromi Arai;O. Fortineau;S. Gambs;Satoshi Hara;Alain Tapp
通讯作者：
U. Aïvodji;Hiromi Arai;O. Fortineau;S. Gambs;Satoshi Hara;Alain Tapp
在可解释的机器学习中评估没有基本事实的解释
DOI：
--
发表时间：
2019
期刊：
arXiv.org
影响因子：
--
作者：
Fan Yang;Mengnan Du;Xia Hu
通讯作者：
Xia Hu