Gordon, John-Stewart

In recent years, numerous researchers on artificial intelligence (AI) and robotics have acknowledged the possibility that a human-level or greater AI may be developed in the coming or even the current century. Against this background, this doctoral dissertation examines the problem of the human moral treatment of greatly sophisticated artificial beings. The central question explored is the following: under what conditions should a robot be regarded as having moral status and rights? In an attempt to answer this question, this dissertation defends the consciousness criterion, which states that the capacity for phenomenal consciousness is necessary for the possession of moral status and rights. In order to substantiate this position, numerous new objections that have been raised against the consciousness criterion in the context of robot rights research are addressed and five distinct arguments are offered against the main competing alternative—the radically relational approach to moral consideration. The research topic is further expanded by providing an analysis on the following issues: the possibility of a hybrid approach to moral consideration, the epistemic problem of detecting consciousness in robots, possible arguments against machine moral considerability, and moral bases for extending legal protection to robots.

The so-called ‘Divine Command Theory’ is the meta-ethical theory which states that God grounds morality ontologically. While it is certainly not a majority view, it has quite a few notable proponents in contemporary analytic philosophy (such as Robert Adams, Paul Copan, William Lane Craig, and James Porter Moreland). The greatest resurgence of the theory happened at the end of the twentieth century when Adams came up with a modified version of the theory, which proved to be compelling to a number of theistic philosophers. The primary goal of this MA thesis is to provide a sophisticated critique of Divine Command Theory. It is argued that the proponents of divine ethics have not been successful in eliminating non-theistic alternatives to their view, and that the Divine Command Theory is not a very compelling view in its own right. In order to make this point, some comparisons will be drawn between Divine Command Theory and meta-ethical naturalism – one of the more standard theories of moral ontology. The first part of the thesis is the theoretical one. It introduces some of the main positions and concepts to be discussed (the most important ones being those of moral objectivity, Divine Command Theory and meta-ethical naturalism). The second part is the analytical one. In this part responses are given to some of the arguments presented by Craig, and some arguments are put forward against Divine Command Theory. Contrary to the claims made by Craig, it is not convincing that on the atheistic worldview human beings would have no objective moral worth because they are ‘nothing special’ (by this it is meant that they are accidental, temporal, and insignificant on the cosmic scale). The non-theistic theories of meta-ethics simply do not posit any of these properties as something that determines one’s moral worth qualitatively. Craig’s evolutionary argument seems to be flawed decisively, whereas his argument from immateriality of human beings is fraught with philosophical controversy. Problems are also pointed out in relation to Craig’s usage of the word ‘special’ in the context of moral philosophy. One argument that is put forward against theistic meta-ethics in this thesis is that it is an intrinsically improbable position, because it necessarily presupposes theism. Aside from that, it seems to have no advantage over other theories in terms of accounting for moral values, moral obligations, and moral epistemology. It is also argued that Divine Command Theory is faced with the problem of arbitrariness. In order to avoid it, the proponents of Divine Command Theory have rejected theological voluntarism and accepted particularistic version of the modified form of the theory. It seems however that this only pushes the problem back one step, because the particularistic model is open to a different kind of arbitrariness objection – that of meta-ethical stopping point.