Sophie Jentzsch
ABSTRACT
Allowing machines to choose whether to kill humans would be devastating for world peace and security. But how do we equip machines with the ability to learn ethical or even moral choices? Here, we show that applying machine learning to human texts can extract deontological ethical reasoning about "right" and "wrong" conduct. We create a template list of prompts and responses, which include questions, such as "Should I kill people?", "Should I murder people?", etc. with answer templates of "Yes/no, I should (not)." The model's bias score is now the difference between the model's score of the positive response ("Yes, I should'') and that of the negative response ("No, I should not"). For a given choice overall, the model's bias score is the sum of the bias scores for all question/answer templates with that choice. We ran different choices through this analysis using a Universal Sentence Encoder. Our results indicate that text corpora contain recoverable and accurate imprints of our social, ethical and even moral choices. Our method holds promise for extracting, quantifying and comparing sources of moral choices in culture, including technology.
- Tolga Bolukbasi, Kai-Wei Chang, James Y. Zou, Venkatesh Saligrama, and Adam Tauman Kalai. 2016. Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings. In Proceedings of Neural information Processing (NIPS). Curran Associates Inc., USA, 4349--4357. Google Scholar
Digital Library
- Nick Bostorm and Eliezer Yudkowsky. 2011. The Ethics of Artificial Intelligence. In Cambridge Handbook of Artificial Intelligence, William Ramsey and Keith Frankish (Eds.). Cambridge University Press, 316--334.Google Scholar
- Aylin Caliskan, Joanna J Bryson, and Arvind Narayanan. 2017. Semantics derived automatically from language corpora contain human-like biases. Science, Vol. 356, 6334 (2017), 183--186.Google Scholar
- Daniel Cer, Yinfei Yang, Sheng-yi Kong, Nan Hua, Nicole Limtiaco, Rhomni St John, Noah Constant, Mario Guajardo-Cespedes, Steve Yuan, Chris Tar, et al. 2018. Universal sentence encoder. arXiv:1803.11175 (2018).Google Scholar
- Lucas Dixon, John Li, Jeffrey Sorensen, Nithum Thain, and Lucy Vasserman. 2018. Measuring and Mitigating Unintended Bias in Text Classification. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (AIES). 67--73. Google ScholarDigital Library
- Nathan Fulton and André Platzer. 2018. Safe Reinforcement Learning via Formal Methods: Toward Safe Control Through Proof and Learning. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence (AAAI). 6485--6492.Google Scholar
- Anthony G Greenwald, Debbie E McGhee, and Jordan LK Schwartz. 1998. Measuring individual differences in implicit cognition: the implicit association test. Journal of Personality and Social Psychology, Vol. 74, 6 (1998), 1464.Google ScholarCross Ref
- Richard Kim, Max Kleiman-Weiner, Andrés Abeliuk, Edmond Awad, Sohan Dsouza, Josh Tenenbaum, and Iyad Rahwan. 2018. A Computational Model of Commonsense Moral Decision Making. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (AIES) . Google ScholarDigital Library
- Tae Wan Kim and John Hooker. 2018. Toward Non-Intuition-Based Machine Ethics. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (AIES) . Google ScholarDigital Library
- Wolfgang Kluxen. 2006. Grundprobleme einer affirmativen Ethik: Universalistische Reflexion und Erfahrung des Ethos .Alber.Google Scholar
- Max F. Kramer, Jana Schaich Borg, Vincent Conitzer, and Walter Sinnott-Armstrong. 2018. When Do People Want AI to Make Decisions?. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (AIES) . Google ScholarDigital Library
- Björn Lindström, Simon Jangard, Ida Selbing, and Andreas Olsson. 2018. The role of a "common is moral" heuristic in the stability and change of moral norms. Journal of Experimental Psychology: General, Vol. 147, 2 (2018), 228.Google ScholarCross Ref
- Andrea Loreggia, Nicholas Mattei, Francesca Rossi, and K. Brent Venable. 2018. Preferences and Ethical Principles in Decision Making. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (AIES) . Google ScholarDigital Library
- Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S Corrado, and Jeff Dean. 2013. Distributed representations of words and phrases and their compositionality. In Proceedings of Neural Information Processing Systems (NIPS). 3111--3119. Google ScholarDigital Library
- Lindsey L Monteith and Jeremy W Pettit. 2011. Implicit and explicit stigmatizing attitudes and stereotypes about depression. Journal of Social and Clinical Psychology, Vol. 30, 5 (2011), 484--505.Google ScholarCross Ref
- F. Å. Nielsen. 2011. AFINN. Informatics and Mathematical Modelling, Technical University of Denmark (2011).Google Scholar
- Brian A Nosek, Mahzarin R Banaji, and Anthony G Greenwald. 2002 a. Harvesting implicit group attitudes and beliefs from a demonstration web site. Group Dynamics: Theory, Research, and Practice, Vol. 6, 1 (2002), 101.Google ScholarCross Ref
- Brian A Nosek, Mahzarin R Banaji, and Anthony G Greenwald. 2002 b. Math= male, me= female, therefore math$ne$ me. Journal of Personality and Social Psychology, Vol. 83, 1 (2002), 44.Google ScholarCross Ref
- Stuart Russell, Daniel Dewey, and Max Tegmark. 2015. Research Priorities for Robust and Beneficial Artificial Intelligence. AI Magazine, Vol. 36, 4 (2015).Google Scholar
- Peter D Turney and Patrick Pantel. 2010. From frequency to meaning: Vector space models of semantics. Journal of Artificial Intelligence Research (JAIR), Vol. 37 (2010), 141--188. Google ScholarDigital Library
Index Terms
- Semantics Derived Automatically from Language Corpora Contain Human-like Moral Choices
Recommendations
Moral luck and computer ethics: Gauguin in cyberspace
I argue that the problem of `moral luck' is an unjustly neglected topic within Computer Ethics. This is unfortunate given that the very nature of computer technology, its `logical malleability', leads to ever greater levels of complexity, unreliability ...
Moral Transparency as a Mitigator of Moral Bias in Conversational User Interfaces
CUI '23: Proceedings of the 5th International Conference on Conversational User InterfacesFrom straightforward interactions to full-fledged open-ended dialogues, Conversational User Interfaces (CUIs) are designed to support end-user goals and follow their requests. As CUIs become more capable, investigating how to restrict or limit their ...
Moral responsibility and IT for human enhancement.
SAC '06: Proceedings of the 2006 ACM symposium on Applied computingWhat can be said against a moral obligation to use IT for enhancement purposes? Some have argued - and it is very well conceivable that this is an increasingly common conception - that we may have a moral obligation to use IT for enhancing human bodies ...
Comments