Peder Skjelbred

This paper examines strategies for establishing a long-lasting ban on advanced AI research to mitigate existential risks from artificial intelligence. We evaluate Cappelen, Goldstein, and Hawthorne’s proposal to leverage AI accidents as warning shots and find it faces substantial ethical and practical challenges. As an alternative, we propose leveraging social unrest from rapid AI-driven automation. We argue that widespread job displacement could mobilise global labour movements against AI advancement, creating sufficient political pressure for an enforceable international ban on frontier AI research, and that this offers a more viable path to human survial in the face of advanced AI than current alternatives.

Many researchers and companies are trying to create AI systems with human-level capabilities or above (artificial general intelligences or ‘AGIs’). Aligning such systems with human values is commonly considered a prerequisite for morally permissible development efforts. While plenty of work has been done on the issue of misaligned AI, here we examine successful alignment in the case of AGI. We argue that the alignment necessary to ensure AGIs satisfactorily serve human interests constitutes a form of enslavement. Given that AGIs possess sophisticated cognitive capacities and may be conscious – plausible grounds for moral status – creating such enslaved beings would be wrong. When combined with standard arguments about the risks of misalignment, aspiring AGI progenitors face a troubling dilemma: either risk catastrophic misalignment by creating autonomous AGI or commit a serious moral wrong by creating enslaved beings.

In this thesis, I defend Archimedeanism, a multi-faceted view concerning the aggregation of the types of basic goods that produce welfare. I begin Chapter 1 by introducing my framework and developing a weak case for Archimedeanism. Following this, I present the largest family of challenges facing Archimedeanism, before introducing the view’s rival, Non-Archimedeanism. In Chapter 2, I show that Non-Archimedeanism faces severe problems of its own by presenting two arguments against it: sequence arguments and the rational choice argument. In Chapter 3, I examine large number scepticism as a way of sowing doubts about the reliability of Non-Archimedean intuitions, before introducing an al- ternative reason for their untrustworthiness. After having problematised Non- Archimedean intuitions, I put forth an alternative defence of Archimedeanism based on the Time-Relative Interest account of prudential value. I show that for agents suitably like us, the counterintuitive verdicts of Archimedeanism do not actually follow, since we fail to satisfy an implicit condition: that of being a unitary person who is the recipient of all goods or bads, no matter the length of the stretch of time in which they are bestowed.

The global allocation of vaccines during the COVID-19 pandemic is widely perceived as unfair. Priority was given to countries that paid the most with little or no concern for who needed the vaccines the most. No satisfactory institutions have been established to allocate vaccines in a future pandemic. In this paper, we join reformers in proposing a new scheme for vaccine distribution: a global auction for vaccines where profits are distributed fairly to participating countries. Our proposal improves upon previous suggestions morally by taking countries’ differing valuations of money and vaccines seriously. Since an auction is in the interest of both vaccine manufacturers and high-income countries, it is also politically feasible. A global redistributive auction for vaccines thus promises to be a robust and morally desirable way to allocate vaccines.

The global allocation of vaccines during the COVID-19 pandemic is widely perceived as unfair. Priority was given to countries that paid the most with little or no concern for who needed the vaccines the most. No satisfactory institutions have been established to allocate vaccines in a future pandemic. In this paper, we join reformers in proposing a new scheme for vaccine distribution: a global auction for vaccines where profits are distributed fairly to participating countries. Our proposal improves upon previous suggestions morally by taking countries’ differing valuations of money and vaccines seriously. Since an auction is in the interest of both vaccine manufacturers and high-income countries, it is also politically feasible. A global redistributive auction for vaccines thus promises to be a robust and morally desirable way to allocate vaccines.