multipolar scenario
In plain wordsA possible future in which many powerful AIs, companies or countries coexist and compete, with none able to dominate the rest.
An outcome of the transition to advanced AI in which power remains distributed among several or many agents, such as states, firms and AI systems of comparable capability, rather than concentrated in a singletonsingleton, n.: A possible future world run by a single decision-maker at the top, such as one government, one alliance or one AI, with no rival able to challenge it.. Multipolarity avoids a single point of failure or tyranny, but analysts identify distinctive dangers: competitive pressure that rewards speed over safety (AI arms raceAI arms race, n.: A competition among countries or companies to build the most powerful AI first, in which fear of falling behind pushes everyone to move faster and cut corners on safety.), erosion of human influence as AI-run institutions outcompete human ones (gradual disempowermentgradual disempowerment, n.: The risk that, as AI does more and more of the work people now do, humanity slowly loses its grip on business, government and culture, until it no longer steers its own future.), wages driven toward the cost of running AI copies, as in Robin Hanson's economics of brain emulations, and conflict among powerful actors. Andrew Critch and others stress multipolar failures arising from interactions among many individually aligned systems, which no single developer's alignment work can prevent.
Even if every lab solves alignment, a multipolar scenario can still go badly.
Origin
Analyzed by Nick Bostrom in Superintelligence (2014), chapter 11, 'Multipolar scenarios', drawing on Robin Hanson's work on emulation economies; extended by Andrew Critch and David Krueger's ARCHES report (2020) and Critch's 2021 essay 'What Multipolar Failure Looks Like'.
Usage note
Related to Scott Alexander's 'multipolar trap', also called Moloch: competition that forces every actor to sacrifice shared values to keep up.
Related
- singletonsingleton, n.: A possible future world run by a single decision-maker at the top, such as one government, one alliance or one AI, with no rival able to challenge it.
- AI arms raceAI arms race, n.: A competition among countries or companies to build the most powerful AI first, in which fear of falling behind pushes everyone to move faster and cut corners on safety.
- gradual disempowermentgradual disempowerment, n.: The risk that, as AI does more and more of the work people now do, humanity slowly loses its grip on business, government and culture, until it no longer steers its own future.
- decisive strategic advantagedecisive strategic advantage, n.: A lead in technology and power so large that whoever holds it, whether a country, a company or an AI, could impose its will on the whole world.
- whole brain emulationwhole brain emulation, n.: Copying a particular person's brain into a computer in such detail, down to individual neurons and their connections, that the simulation would think and behave like that person.
- slow takeoffslow takeoff, n.: A scenario in which AI becomes superhuman step by step, over years or decades, as it spreads through the economy, giving society time to spot problems and respond.
- singletonsingleton, n.: A possible future world run by a single decision-maker at the top, such as one government, one alliance or one AI, with no rival able to challenge it.
Mentioned in
AI arms raceAI arms race, n.: A competition among countries or companies to build the most powerful AI first, in which fear of falling behind pushes everyone to move faster and cut corners on safety., gradual disempowermentgradual disempowerment, n.: The risk that, as AI does more and more of the work people now do, humanity slowly loses its grip on business, government and culture, until it no longer steers its own future., slow takeoffslow takeoff, n.: A scenario in which AI becomes superhuman step by step, over years or decades, as it spreads through the economy, giving society time to spot problems and respond., takeofftakeoff, n.: The stretch of time between AI reaching roughly human level and AI becoming vastly superhuman. People argue over whether it would take decades, years, months or just days., decisive strategic advantagedecisive strategic advantage, n.: A lead in technology and power so large that whoever holds it, whether a country, a company or an AI, could impose its will on the whole world., Robin HansonRobin Hanson, prop. n.: An American economist who debated Eliezer Yudkowsky about whether AI would take off suddenly or gradually, and who imagined an economy run by digital copies of human minds., singletonsingleton, n.: A possible future world run by a single decision-maker at the top, such as one government, one alliance or one AI, with no rival able to challenge it., whole brain emulationwhole brain emulation, n.: Copying a particular person's brain into a computer in such detail, down to individual neurons and their connections, that the simulation would think and behave like that person.