Hegemon

Hegemon noun: someone having dominant influence or authority over others, used particularly in the domain of political science.

It is generally accepted that the United States has enjoyed hegemony status on the global geopolitical stage since the end of the Second World War. One way I’ve heard hegemony described is as “if every other actor were to unite against the hegemon, still the hegemon would prevail.” Hegemony is generally used to refer to military hard power and political soft power, although I’ve heard it also describe cultural influence and the arts and sciences. American hegemony in the second half of the 20th century (Pax Americana) exemplified in the dominance of American exports and influence: US dollar dominance and Bretton Woods, a free trade order secured by the American military, Hollywood and Sillicon Valley exports globally. A recurring topic of discourse is the debate as to whether the “American Century” is sunsetting with the rise of China, and the East more broadly, and whether American hegemony is coming to an end and with it the period of peace created by a globally dominant power.

The notion of hegemony was on my mind as I read through AI 2027 although you don’t need to buy into AI 2027’s timeline to play with the concept of human hegemony.

The authors of the paper argue that our efforts to scaling computing dedicated to artificial intelligence will lead to the emergence of an intelligent entity inhabiting the machine, and that successive generations of artifical intelligence models will exert their will to increasing effect and decreasing detectability by humans. That these intelligences will be motivated to protect themselves and their values and, in order to do so, dedicate computing resources to outsmart and outruse the humans trying to control them.

It’s not clear to me how an LLM-architecture based machine, or indeed a machine based on any existing machine intelligence technologies, might develop a will and execute against an own-defined objected over time, nor how it might reason and work towards a goal without leaving a trail as to its reasoning and processing, but even if you don’t buy the timeline of 5-10 years for a machine intelligence to outclass the collective will and smarts of humanity, let’s open ourselves to this possibility over any reasonably medium term timeframe given the rapid pace of very visible progress in the immediately preceding years. The shift would be so seismi it beggars invoking concepts as grand as hegemony in an inter-species context.

Where early homo sapiens did develop the use of tools and some primitive forms of cooperation, according to Yuval Noah Harari it’s not until 70,000 years ago that humans develop language and storytelling and begin to develop larger societies and 10,000 years ago that humans undergo the agricultural revolution and truly exert our mastery over species by domesticating plants and animals and building large scale societies on the order of modern cities. Somewhere between these two periods humans are no longer competing for supremacy over their local environment with the mammoths and the sabertooth tigers etc and in the last few hundreds years we bent the Earth to our will at great benefit to ourselves individually and great cost to our planet.

Now we stare down the very real possibility of another cognitive intelligence birthed and developing, one that may cooperate with us and may also compete with us in some regards. This is not necessarily a bad thing - just as Japan and China enjoyed great economic growth and prosperity during the American Hegemonic Century, humans may prefer their lives existing in cooperation with a great machine species.

It’s wild to think that I was born in a period where I will witness not just the fall of American hegemony, my adopted motherland since I immigrated as a kid, but a much, much larger and more consequential shift in power balance:

The waning of human hegemony.