Problems with “Friendly AI”

Ethics and Information Technology 23 (3):543-550 (2021)
  Copy   BIBTEX


On virtue ethical grounds, Barbro Fröding and Martin Peterson recently recommended that near-future AIs should be developed as ‘Friendly AI’. AI in social interaction with humans should be programmed such that they mimic aspects of human friendship. While it is a reasonable goal to implement AI systems interacting with humans as Friendly AI, I identify four issues that need to be addressed concerning Friendly AI with Fröding’s and Peterson’s understanding of Friendly AI as a starting point. In a first step, I briefly recapitulate Fröding’s and Peterson’s arguments for Friendly AI. I then highlight some issues with Fröding’s and Peterson’s approach and line of reasoning and identify four problems related to the notion of Friendly AI, which all pertain to the role and need for humans’ moral development. These are that one should consider the moral tendencies and preferences of the humans interacting with a friendly AI, it needs to be considered whether the humans interacting with a Friendly AI are still developing their virtues and character traits, the indirect effects of replacing humans with Friendly AI should be considered with respect to the possibilities for humans to develop their moral virtues and that the question whether the AI is perceived as some form of Artificial General Intelligence cannot be neglected. In conclusion, I argue that all of these four problems are related to humans moral development and that this observation strongly emphasizes the role and need for humans moral development in correlation to the accelerating development of AI-systems.



    Upload a copy of this work     Papers currently archived: 76,503

External links

Setup an account with your affiliations in order to access resources via your University's proxy server

Through your library

Similar books and articles

Friendly atheism revisited.William L. Rowe - 2010 - International Journal for Philosophy of Religion 68 (1-3):7-13.
Friendly AI.Barbro Fröding & Martin Peterson - 2020 - Ethics and Information Technology 23 (3):207-214.
How to Be a Friendly Skeptical Theist.Francis Jonbäck - 2012 - Forum Philosophicum: International Journal for Philosophy 17 (2):197-210.
How to Be a Friendly Skeptical Theist.Francis Jonbäck - 2012 - Forum Philosophicum: International Journal for Philosophy 17 (2):197-210.
Toward Eco-Friendly Aesthetics.Sheila Lintott - 2006 - Environmental Ethics 28 (1):57-76.
How Are Moral Foundations Associated with Climate-Friendly Consumption?Annukka Vainio & Jaana-Piia Mäkiniemi - 2016 - Journal of Agricultural and Environmental Ethics 29 (2):265-283.
Sympathy, Impartiality, and Care.Susan V. H. Castro - 2017 - Southwest Philosophy Review 33 (2):69-76.
Intolerance and Argument Expression.Ben Cross - 2019 - Social Theory and Practice 45 (3):329-352.
A defense of Frankfurt-friendly libertarianism.David Widerker - 2009 - Philosophical Explorations 12 (2):87 – 108.


Added to PP

20 (#566,467)

6 months
3 (#228,256)

Historical graph of downloads
How can I increase my downloads?