On an LLM Revolution

With the rise of AI and automated cognition, we’re facing a new technological revolution.

Rapid technological change has the potential to radically shift existing power structures. The outcome of this radical potential is dependent on institutions and existing power structures around the technology. This is apparent again and again throughout history. The book “Why Nations Fail” by Daron Acemoglu and James Robinson showed how the poor became rich through a steam power revolution. When the Industrial Revolution came to Britain, it rapidly overthrew the existing class structures. These technologies were not nationalized under the aristocracy or controlled by a small elite. They were new. And so Britain urbanized, concentrated populations in cities, and industrialized by building railways and steamships, and a new class of wealth emerged from merchants, traders, and industrialists. For countries where the technology could be adopted by many, where the technology became an armament of the people rather than the elite, the aristocracy, or the dictatorship, in these nations technology served to empower the people and their broad goals. Power structures shifted tremendously. In the end, this shift of wealth enabled the political enfranchisement of more of the British nation. The first Reform Act of 1832 generally increased enfranchisement of the working class, affording most property owners voting rights. The Chartist movement increasingly enfranchised non-property owning masses from 1832 until the 20th century. By 1918 women began to gain voting rights.

At the same time, this technology was encountered by the Austro-Hungarian Empire ruled by Francis I (also known as Francis II, confusingly) until 1835. During this time, the state struggled to maintain control over its people, as one of the last strongholds of absolute rule still clinging to feudal roots. So when encountering the proposal of a railway in Austro-Hungary, Francis I put it bluntly: “No, no, I will have nothing to do with it, lest the revolution might come into the country.” Further, Austro-Hungarian officials were quoted, “We do not desire at all that the great masses shall become well off and independent… How could we rule over them otherwise?” The spirals of history have not yet revealed the role of technology in our collective fate. But these quotes reveal that broadening technology to the people can broaden our ethics. And the power of changing societal values deeply upsets the narrow ethics of the elite. We hear worried echoes today. We hear them in China where access to the Internet is tightly gated by the state’s technological dominance. We hear them in the United States when Amazon worker unions are continuously suppressed. We hear them in widespread Big Tech antitrust cases. The ethics of the people, enabled by the wealth of the people, enacted by technology made by people and for people, is not in the interest of concentrated power. In the past, in Britain, technology shifted wealth, broke existing class systems, and empowered the masses; in nations where the political landscape allowed for technological oppression and power concentration, society’s goals were much narrower, determined by aristocracies and elites to serve power consolidation. Technological change in Britain transferred power to the people and broadened society’s goals.

The same is happening with AI.

Countries where the people are poised to understand and make use of the technology, within the bounds of safe and non-self-improving usages, will have widespread prosperity. Those where the technology is understood by few and accessible to fewer will enrich only the few. It will focus on wealth concentration, not sustaining life.

That’s not to say AI technology should be unregulated or that it does not pose threats. It is to say that AI is most dangerous when wielded by few and the incentives are to concentrate wealth and power rather than to sustain life for many. There should be regulations – on the scale of companies and nations that are capable of oppressing the people. And where there are local peoples who could create more broad intelligence there should be more distributed technological power – more AI.

Everyone thinks they’re moral. Everyone is in line with their own idea of morality. To align the goals of AI to support life on Earth, we need to broaden the ethics we cultivate through the power of technology. More people need to develop technology and its ethics for themselves, the environment, nature, and the rest of life on Earth that allow us to live and breathe here.