Journal of Artificial Intelligence and Consciousness

“The Prospect of a Humanitarian Artificial Intelligence: Agency and Value Alignment” reunites fields of inquiry that are typically kept separate-philosophy of mind, theory of knowledge, psychology, the social and political sciences, law and ethics. What we have learned to call “artificial intelligence” (AI), and the ways we may use it — intelligently — to promote life, knowledge, freedom, care, altruism, is a technology that must be democratized if it is to be used properly. But we have to be cautious to call it “intelligent” before seeing more carefully how it affects us, humans as a whole.

This paper critically examines Carlos Montemayor’s human rights-based approach to AI ethics, as presented in his 2023 book The Prospect of a Humanitarian Artificial Intelligence. Montemayor proposes that the concept of human rights, grounded in the cognitive needs of human beings, should guide AI development. This paper challenges Montemayor’s moral value-realism (the belief in objective moral values) by arguing for a more inclusive ethical framework that reconciles moral value-realism with skepticism. By proposing an agnostic stance towards moral truths, it suggests a new framework for AI ethics based on subjective intrinsic valuations and shared moral commitments, rather than on objective moral truths. In this proposed framework, without either denying or asserting the existence of absolute values and moral truths, human rights in AI ethics are grounded in collective moral commitments to value human beings intrinsically, rather than in inherent moral truths. This facilitates a more pragmatic and inclusive solution that allows for a diversity of ethical and metaethical perspectives while aligning with the complex nature of AI’s societal integration. The paper concludes that while Montemayor’s work is foundational, it invites further adaptation to develop AI systems that are ethically sound and in harmony with a broad spectrum of human values.

In his book The Prospect of a Humanitarian Artificial Intelligence (2023), Carlos Montemayor proposes, among other things, an account of genuine artificial intelligence (IA) in terms of the notion of attention. This notion, when disconnected from problematic notions like phenomenal consciousness, can account for epistemic agency, which is a central feature of intelligent behavior. This paper questions, on the one hand, whether attention really suffices for epistemic agency and for AI. On the other hand, it is also argued that we cannot completely disentangle phenomenal consciousness from questions about the nature of intelligence, epistemic agency, and attention. After all, we may need to deal with deeper metaphysical questions.

The Prospect of Humanitarian Artificial Intelligence (The Prospect of HAI) encapsulates the vast and intricate landscape of Artificial General Intelligence (AGI) with both theoretical and practical nuances. Carlos Montemayor (Carlos) artfully paints a complex picture, presenting a realm of profound philosophical interest. While abstract, the book navigates the terrain crucial for AGI development, successfully capturing its multifaceted nature. This review focuses on the alignment problem, probing its epistemic and moral dimensions. Imperfect as we are, aligning AGI presents challenges: discerning viable solutions remains uncertain. The epistemic aspect of alignment of the AGI seems to be impossible for the goals Carlos is proposing for the nascent industry of AI. The moral, political, and legal aspects of alignment offer a challenging panorama, mirroring parallels between the Philosophy of AI and theological discourse. Drawing parallels with our experience in moral alignment, especially in guiding children, underscores the idea that this problem is older and more widespread. The Prospect of HAI serves as an inaugural exploration in this realm. Future endeavors should delve deeper into technical intricacies while maintaining a broad philosophical scope. It stands as a compelling precursor to the burgeoning field of philosophy of AI. Carlos should go for the 2nd edition.

In this article, we critically examine the implementation of what Carlos Montemayor calls Humanitarian Artificial Intelligence from a decolonizing perspective. We highlight the dichotomy of optimism and fear surrounding AI, elucidating its potential to address fundamental human problems and the risks of monopolistic control. We critique Montemayor’s proposal to align AI with a human rights framework, arguing that it insufficiently addresses global inequalities. Our tripartite analysis focuses on the distribution of AI resources, language inclusion, and content diversity to ensure AI benefits all humanity. We emphasize the need for equitable access to AI, linguistic diversity in AI training data, and the preservation of marginalized epistemologies. We advocate for strategies to mitigate environmental impacts and avoid cultural imperialism disguised as altruism, calling for a balanced approach between private sector innovation and state regulation to foster a truly humanitarian AI.

This short paper traces some convergences between the ontologies of mind and AI presented in Carlos Montemayor’s Precis for The Prospect of Humanitarian Artificial Intelligence and Enactivism, with a focus on autonomy, embeddedness, perspective, and social interaction. I argue that it is surprising that Montemayor does not engage with any of the traditional 4e ontologies here, especially enactivism, considering his strong reliance on biology and autonomy in carving up intelligence and consciousness.