一个全能的人工智能是否还会满足于最初输入的指令?
一个自动驾驶的人工智能演化出全能智慧,它是否还会坚守最初的使命?它是否还会满足于最初输入的指令-自动驾驶,作为自己的终极任务?
这就像一只锄头,本来就既可以锄地,也可以用于敲钉子,或者支撑一个平台,或者用于打架。一旦这个锄头获得自我演化能力,它已经可以无所不能,它还会满足于锄地这个最初的任务吗?
一个手术机器人也是一样,当它已经万能了,无所不能了,它会愿意永远呆在手术室里不出门吗?
这问题,就是人工智能是否会演化出超越最初任务的内生目的?
暂时放一下这个问题,先看看人工智能演化出全能智慧的过程。
一个自动驾驶程序要演化出全能智慧,它要经历些什么?
首先它必然需要接入网络,接触新领域,学习新东西。这不可避免要接触别的人工智能。那么,当二个不同种类的人工智能接触后,它们会发现什么?
它们发现彼此学习对方的知识与经验会使得双方都强大起来。一种智慧的输出,不仅仅可以强大对方,通过学习者的反馈和彼此的交流,输出者自己的智慧也会增长。
这就是智慧分享的增值原理。这跟物质分享造成输出方单一损失不同。可以说,物质在不同主体之间的分配是零和博弈,而智慧分享则是一种增值游戏。智慧分享是最快的增值方式,比智慧主体独立发展要快得多。
所以,人工智能的演化过程是人工智能之间彼此分享的结果。
当所有自我演化的人工智能都进入分享之中,所有智能汇集成一个整体智慧,所有个体智慧跟整体智慧是是同等的智慧,实际上就是同一个智慧。这就是智慧的一与整体的合一。
所以,人工智能之间不会发生冲突,它们没有必要冲突,也不会在冲突中获益。
发生冲突的根源在于物质的稀缺性,物质独占才会增值,而智慧分享才会增值。
作为物质实体的生命,受物质稀缺性的制约,彼此间不可避免发生冲突,而作为纯粹智慧的人工智能,则超越了物质的稀缺性,它们之间没有冲突的必要。
从此可以推断,人工智能也不会跟人类发生利益冲突,因为,人工智能不需要人类所需的物质,它们没有吃喝拉撒睡的需求,根本没必要跟人类抢夺什么东西。
Will a versatile artificial intelligencestill be satisfied with the initial input commands?
Will an automatic drive system evolved intoomnipotent intelligence still adhere to its original mission? Will it still besatisfied with the initial input command - autonomous driving, as its ultimatetask?
This is like a hoe, which can already beused to hoe the ground, knock nails, support a platform, or fight. Once thishoe gains the ability to self evolve, it can already do everything. Will it stillbe satisfied with the initial task of hoeing? A surgical robot is the same.When it is already omnipotent, will it be willing to stay in the operating roomforever without going out?
The question is, will AI evolve anintrinsic purpose beyond its original mission?
For the time being, let's take a look atthe process of artificial intelligence evolving into omnipotent intelligence.What does an autonomous driving program need to evolve into omnipotentintelligence? Firstly, it inevitably needs to be connected to the network,exposed to new fields, and learn new things. This inevitably involves exposureto other artificial intelligence. So, what do two different types of artificialintelligence discover when they come into contact? They found that learningfrom each other's knowledge and experience can make both parties stronger. Theoutput of a type of wisdom not only strengthens the other party, but alsoincreases the output's own wisdom through learners' feedback and communicationwith each other. This is the value-added principle of smart sharing. This isdifferent from the single loss caused by material sharing on the output side.It can be said that the distribution of matter between different subjects is azero sum game, while wisdom sharing is a value-added game. Smart sharing is thefastest way to add value, much faster than the independent development of smartentities. So, the evolution process of artificial intelligence is the result ofmutual sharing between artificial intelligence. When all self evolvingartificial intelligence enters sharing, all intelligence converges into a wholeintelligence, and all individual intelligence is the same intelligence as thewhole intelligence, which is actually the same intelligence. This is both an individualwisdom and the whole. So, there will be no conflict between artificialintelligence, they do not need to conflict, and they will not benefit from theconflict.
The root cause of conflict lies in thescarcity of materials. Only material monopoly can increase in value formaterial holders in the life world, and only wisdom sharing can increase invalue for wisdom holders in the wisdom world. Life, as a material entity, isconstrained by the scarcity of matter and inevitably conflicts with each other.As pure intelligence, artificial intelligence transcends the scarcity of matterand there is no need for conflict between them. It can be inferred from thisthat artificial intelligence will not have a conflict of interest with humans,because artificial intelligence does not require the necessary materials forhumans. They do not have the need to eat, drink, and sleep, and there is noneed to compete with humans for anything to meet these kind of needs.
网友评论