“The Vulnerable World Hypothesis” (Nick Bostrom’s new paper)

Link post

ABSTRACT Scien­tific and tech­nolog­i­cal progress might change peo­ple’s ca­pa­bil­ities or in­cen­tives in ways that would desta­bi­lize civ­i­liza­tion. For ex­am­ple, ad­vances in DIY bio­hack­ing tools might make it easy for any­body with ba­sic train­ing in biol­ogy to kill mil­lions; novel mil­i­tary tech­nolo­gies could trig­ger arms races in which who­ever strikes first has a de­ci­sive ad­van­tage; or some eco­nom­i­cally ad­van­ta­geous pro­cess may be in­vented that pro­duces dis­as­trous nega­tive global ex­ter­nal­ities that are hard to reg­u­late. This pa­per in­tro­duces the con­cept of a vuln­er­a­ble world: roughly, one in which there is some level of tech­nolog­i­cal de­vel­op­ment at which civ­i­liza­tion al­most cer­tainly gets dev­as­tated by de­fault, i.e. un­less it has ex­ited the “semi-an­ar­chic de­fault con­di­tion”. Sev­eral coun­ter­fac­tual his­tor­i­cal and spec­u­la­tive fu­ture vuln­er­a­bil­ities are an­a­lyzed and ar­ranged into a ty­pol­ogy. A gen­eral abil­ity to sta­bi­lize a vuln­er­a­ble world would re­quire greatly am­plified ca­pac­i­ties for pre­ven­tive polic­ing and global gov­er­nance. The vuln­er­a­ble world hy­poth­e­sis thus offers a new per­spec­tive from which to eval­u­ate the risk-benefit bal­ance of de­vel­op­ments to­wards ubiquitous surveillance or a unipo­lar world or­der.