AI and Human Responsibility

As we move further into a world in which artificial intelligence is a part of everyday life, we must consider the psychological effects this technology is having on us individually and collectively. We have to keep asking ourselves what we want this technology to be, how we want to shape it and how we are willing to be shaped by it. 

There is a fight going on currently between the United States government and Anthropic, one of the major AI tech companies. Though the debate centers on the government's ability to use Anthropics AI technology in its military operations, the outcome of this fight will have major implications for the relationship between humanity and artificial intelligence. 

There is a contingency within the government that wants to fast-track the implementation of AI into military operations. This contingency, made up of government officials and military officers, view AI as a way to more effectively deliver violence. Such thinking reveals an underlying belief in the inevitability of violence, and perhaps in their mind, the worthwhileness of violence as an objective. 

The military, like other organizations grappling with AI, must give significant thought to the potential human cost of developing AI without regard for the psychological effects this technology is having on individuals and communities. Some people think AI is dangerous, and this view may be justified as there does appear to be a dark and destructive vision of AI forming in some segments of the government. Military officials who support this vision view AI as a tool that can enhance and augment the ability to inflict mass destruction without being hampered by ideological constraints.

I’m struck by that phrasing—without ideological constraints—struck by how disingenuous they are. Ideologies are simply systems of ideas, they are a specific manner of thinking, and even the most vague and ill-defined patterns of thought can represent some form of them.

The dream of an AI that allows someone to operate without ideological constraint is a contradiction that doubles as an admission of a specific kind of ideology, one whose guiding principle is action without inhibition. If used in this way, AI will become a tool that further cements a pattern of relating amongst groups that hinges on their being a dominant and a subordinate party. This inevitably leads to more conflict and turns any notion of AI being able to usher in a new era of prosperity and flourishing into nothing more than a clever marketing ploy. 

When it comes to the debate over the use of AI in the military, two camps seem to be emerging, which I will respectively call the functionalists and the aggressionists.

Functionalists seem to advocate for AI that can reduce the cognitive load placed on military servicemen so that they can focus on tasks that are more central to their missions. Functionalists believe that human operators should still be the ones making crucial decisions. 

Aggressionists share some of the same goals as the functionalists, but they are also more bold in their ambitions. This group is more likely to view the development of AI as a new Manhattan Project. Beyond improved functionality, what they are most concerned about is winning what they believe is the new global arms race, with AI as the deciding factor. 

Thus far the aggressionist viewpoint seems to be winning out. The powers of the judiciary are currently being turned on Anthropic and used as leverage in their dispute with the government. Vague and innocuous phrases like “supply chain risk” are being used to brand Anthropic as a serious threat to national security. It is difficult to make the CEO of a multi-billion dollar corporation into a sympathetic figure, but that’s the point we’ve somehow arrived at in this saga. 

The view that unfettered access to AI is crucial to military success is complicated by the fact that success is being defined as the ability to create lethal effects at scale. We should all be leery of following such logic.

AI should be developed in ways that serve the interests of humanity and produce the most positive good. Such outcomes are only possible if we advance beyond narrowly thinking about AI only in terms of military power.