John Carter (St. Thomas University) has posted Algorethics and Neuro-Symbolic Military AI: Ontologies, “Hellscapes,” Just War, and Silicon Valley on SSRN. Here is the abstract:
Algorethics, as described in the Rome Call for AI Ethics (2020), is an overarching field that includes algorithmic accountability, systemic AI governance, value-sensitive design (VSD), responsible AI (RAI), trustworthy AI, explainable AI (XAI), and similar terms, but which puts these subfields in conversation with broader, more historical ethical and moral traditions. This essay evaluates the ways in which ontology-supported neuro-symbolic artificial intelligence (NeSy) addresses algorethics concerns about explainability and transparency in military applications of AI but does not solve the problem of accountability, which in the case of AI may be greater than similar concerns regarding other private defense contractors. Existing conflicts in Ukraine and Sudan and a potential “hellscape” strategy in the Taiwan Strait are discussed, as are recent U.S. military actions in Venezuela and Iran, and the crossover effects among them based on common suppliers of AI-supported drones; in Sudan and parts of Ukraine, the label “hellscape” is already apt. A parallel is drawn between John Ford, SJ’s condemnation of obliteration bombing in World War II and the drone-swarmed “hellscapes” that current military technology makes possible. The dispute between Anthropic and the U.S. Defense Department over usage restrictions is discussed, as is the role of Palantir and its stated commitment to norms of just war. In addition to Ford, the work of Francisco de Vitoria and Alasdair MacIntyre is also brought to bear on the problem.