Virtuous Artificial Intelligence 公开
Goldfein, Benjamin P. (Spring 2018)
Published
Abstract
My aim in this project is to challenge the consequentialist narrative that morally-sentient artificial intelligence (AI) should be machine-utilitarians. I begin by asserting that we must understand how to conceive AI before we can sketch a blueprint of a future with person and non-person morally-sentient beings. I then argue that due to recent explosive advancements in AI, persons’ and AIs’ moral sentiences will soon be indistinguishable. When this happens, AIs should abide by a system of ethics to ensure the protection of person and non-person morally-sentient beings. Furthermore, I assert that ethical morally-sentient AIs would necessarily follow a system of Aristotelean virtue ethics. Following this ethical system would equip AIs to balance competing obligations through creative ‘judgment calls’ and corrective prohairesis. I conclude by showing how an AI virtue ethicist would be able to learn prima facie virtues, thus removing the supererogatory burden off of the programmer to code the ‘perfectly-ethical’ AI.
Table of Contents
Foreword ............................................................ 1
Defining Artificial Intelligence ............................ 5
Three Categories ............................................. 6
Artificial Narrow Intelligence ....................... 6
Artificial General Intelligence ....................... 9
Artificial Superintelligence .......................... 11
Regarding Moral Sentience ................................ 13
Thinking Artificial Intelligence ............................. 18
Functionalism ........................................... 19
Turing Test .............................................. 21
Searle’s Chinese Room .............................. 24
Further Thoughts ...................................... 27
Could Have Done Otherwise ............................. 28
Setting the Stage ...................................... 29
Problem of Freewill .................................... 31
Determinism and Indeterminism .................. 32
Moral Capacity and Potential ....................... 35
Applied to Artificial Intelligence .................... 38
Considering Ethical Systems ............................... 41
Consequentialism ............................................ 42
Terminology .............................................. 44
Utilitarianism ............................................. 47
Utilitarian Artificial Intelligence ..................... 49
Aristotelean Virtue Ethics ................................... 53
Overarching Principles ................................. 54
Corrective Prohairesis .................................. 57
Artificial Intelligence Virtue Ethicists .............. 59
Afterword ............................................................ 63
Bibbliography ...................................................... 64
About this Honors Thesis
School | |
---|---|
Department | |
Degree | |
Submission | |
Language |
|
Research Field | |
关键词 | |
Committee Chair / Thesis Advisor | |
Committee Members |
Primary PDF
Thumbnail | Title | Date Uploaded | Actions |
---|---|---|---|
|
Virtuous Artificial Intelligence () | 2018-04-09 14:57:18 -0400 |
|
Supplemental Files
Thumbnail | Title | Date Uploaded | Actions |
---|