The increased prevalence of AI technology developed or adapted for military use raises difficult questions relating to the compatibility of this new technology with international law in general, and international human rights law (IHRL) in particular. This is because the move away from human discretion and agency to decision making by machines in contexts involving the use of lethal force implicates some of the most basic human rights, including the right to life. Indeed, the Human Rights Committee, the expert body entrusted with monitoring the application of the International Covenant on Civil and Political Rights, has expressed in 2018 its view on the relationship between the emergence of new military AI and respect for the right to life.

The present article will review the terms of the IHRL debate surrounding the introduction of AI technology into military contexts and its relationship to the right to life. In Part One, I will review briefly some actual and potential applications of AI in military contexts. Part Two deals with three principal objections against introducing military AI to battlefield environments: the capacity of autonomous or semi-autonomous AI systems to properly apply international humanitarian law (IHL), concerns about de facto lowering of standards of humanitarian protection and other ethical and legal implications of transferring certain life and death decision from humans to machines. Part Three reviews, in light of the three principled objections, specific proposals by the ICRC to limit the use of AI in military contexts (limiting the scope and manner of use of autonomous weapon systems, and excluding unpredictable and lethal systems). And Part Four reviews the main issues discussed in this article from the vantage point of the right to life under IHRL, as elaborated in General Comment No. 36. 

The paper uses aspirations mentioned in the initial research on machine learning decentralization as a lens for examining the cur- rent state-of-the-art and exposing opportunities for future inno- vations. We explore the potential and limitations of decentralized architectures in affording privacy and human agency for end users, competition, and collaboration for wider market and civic players. We then elaborate on the legal and technological developments necessary for decentralized machine learning systems to realize their liberating potential.

The paper uses aspirations mentioned in the initial research on machine learning decentralization as a lens for examining the cur- rent state-of-the-art and exposing opportunities for future inno- vations. We explore the potential and limitations of decentralized architectures in affording privacy and human agency for end users, competition, and collaboration for wider market and civic players. We then elaborate on the legal and technological developments necessary for decentralized machine learning systems to realize their liberating potential.