The Coming War Over The Technosphere

The first threat is dangerous artifacts produced by individuals, and the second is war between groups that see technological progress as the threat, and those who see it as the solution

Every day, you are assaulted by billions of agents trying to eat you from the inside out. Over billions of years, microbes evolved to invade organisms and reproduce. In response, complex lifeforms developed immune systems to counter them. The struggle of individual organisms emerges as an equilibrium of the entire ecosystem. In the next few decades, intelligently designed agents will become practical and widely accessible thanks to AI tools, forcing humanity to adapt.

If Life 1.0 is the biosphere, Life 2.0 is the memesphere, or the struggle over culture and ideas of the higher order animals and human societies. Ideas have their own evolutionary struggle, as they enable their hosts to adapt to environments much faster than the relatively slow pace of genetics.

Life 3.0 is the technosphere. In 2020, the total mass of the materials created by humanity (anthropogenic mass) exceeded the total mass (biomass) of all living things on Earth -- about 1.1 teratones. https://www.nature.com/articles/s41586-020-3010-5

The technosphere is mostly inanimate -- concrete, metal, plastics, etc. Houses, bridges, tools and appliances are mainly a distilled and frozen form of the memesphere. But since the mid 20th century, the technosphere increasingly replicates human thinking. Automatons were created to recreate thought to save human labor. Modern vehicles and airplanes can self-optimize operating parameters and employ a variety of algorithms to avoid accidents. This is a frozen form of human thought, but AI systems are second-order automatons that not only reproduce algorithms but write their own.

Second order automations like large language models vastly expand the power of individuals and groups to influence the environment because the people involved no longer have to understand every single algorithm programmed into the models. A single operator can now leverage vast fields of knowledge without any understanding of their contents -- they just need to source raw materials and follow the instructions of the AI.

The vast majority of AI applications will surely be tremendously beneficial for humanity, but the great threat is from individuals who maliciously, negligently, or innocently release synthetic agents into the technosphere that can do great harm.

Computer viruses spread through early information networks, but we built robust immune systems in response, and antivirus software on individual desktops is now almost entirely redundant. Soon, organic and non-organic physical agents will be widely produced at the instruction of AI systems, and the technosphere will need to develop a robust material immune system.

I predict that this will not happen without a number of serious disasters to underscore the need for technosphere immune systems. In the struggle between offensive and defensive tools, there is often a temporary gap that favors one side - a technological overhang. Because humans are not good at responding to new threat paradigms, a technological overhang will initially develop for the offense, which will only be addressed after a series of disasters -- which could scale from infrastructure destruction to millions or billions dead from engineered viruses. The best we can plan for is to recognize and mitigate the threat before casualties mount.

While some AI doomers predict a single existential threat that arises with the first superintelligence, a more plausible threat is billions of individual humans, all with the power for good and evil. Individually, they might not be able to pose an existential threat, but in aggregate, their potential to create dangerous biological or mechanical agents is still a major threat. The threat doesn't have to be nanobots: engineered plants and animals combine 4 billion years of evolution with intelligent design. An unstable locust swarm can be just as much of an existential threat as an engineered virus or grey goo.

In response, the technosphere will develop a defense in depth: every human being will circulate intelligent agents in their bloodstream that filter out dangerous biological agents, including evolved ones like the common cold as well as nanobots. Each home will contain a HEPA filter that also neutralizes nanobots. In space, satellites and drones will scan for technosignatures of dangerous swarms and deploy defensive swarms in response. (Neal Stephenson's book The Diamond Age imagines how some of this could work.)

In the next few decades, humanity will face two conflicts: the first with dangerous technological artifacts produced by malicious or incompetent individuals, and the second between groups that see technological progress as the threat, and those who see it as the solution. Both threats pose an existential threat to intelligent life in the universe. Hugo de Garis calls these sides the "Terrans" and "Cosmists" in his book "The Artilect War." The technosphere war will be an intellectual conflict between humans, as a consensus emerges about the best way to develop an immune system against malicious uses of advanced technology.