SOPhiA 2021

Salzburgiense Concilium Omnibus Philosophis Analyticis

SOPhiA ToolsDE-pageEN-page

Programme - Talk

On the Possibility of Assigning the Full Moral Status (FMS) to the Artificial General Intelligent (AGI) Systems
(Ethics, )

A With due course of time, the AI system has become more and more flexible and adaptable and can perform a large variety of cognitive tasks. Simultaneously, the capacity of AI systems to act autonomously has also been enhanced extensively. At present, AI systems can identify objects from video and image; transcribe speech and translate between languages, Stock trading; drive automobiles; fly drones; write its encryption language; diagnose cancer in tissues, and so on. The modern AI has won against humans at Jeopardy, and at Go. IBM's Watson has defeated two best human players Ken Jennings and Brad Rutter at Jeopardy in 2011. DeepMind's AlphaGo, which is based on an advanced search tree and deep neural networks has won against the three-time European Go Champion Fan Hui in 2015 and eighteen times world Go champion Lee Sedol in 2016. Over time, DeepMind has introduced an extended version of AlphaGo called AlphaGo Zero, which can learn the game by playing it. In late 2017, DeepMind has again launched a new version of the AI called AlphaZero, a self-taught AI algorithm that has become a master in chess, Shogi, and Go. The latest version of DeepMind's AI algorithm, which is called MuZero has taken AlphaZero's ideas one step further and can play Go, Chess, Shogi, and Atari games without knowing any rules of these games. Since AI technologies have continued to move forward and have become more and more capable over time, the question about the ethics of artificial intelligence has become more vital than ever. When philosophers are concerned with the moral status of AI systems, the focus of such status is expected to fall on AGI systems. In the long term, AI researchers and futurists have anticipated developing AGI that can perform any intellectual task similar to a human being. Moreover, the development of such intelligent systems in the future may give rise to the question of the moral status of such systems, i.e. whether it is possible to assign them full moral status (FMS) or not. To investigate the possibility, I put forward an argument, which states that if two entities (a human being and an AGI system) have similar functionality and similar sophisticated cognitive capacity, but they differ in the substrate of their implementation and ontogeny, then they have similar moral status. In conclusion, I argue that the AGI systems may have full moral status (FMS) by their similar functionality and sophisticated cognitive capacity as normal human beings.

Chair: Francesca Miccoli
Time: 16:00-16:30, 11 September 2021 (Saturday)
Location: SR 1.005
Remark: (Online Talk)

Mubarak Hussain  
(Indian Institute of Technology Dharwad , )

Testability and Meaning deco