By Ashoke Agarrwal
I find it ridiculous that a school of learned and credentialed humans fear that if humans are not careful someday, not in the distant future, AIs will take over the world and destroy human civilisation. The assumption underlying that stream of thinking is that Artificial Intelligence (AI) systems, as they progress towards Artificial General Intelligence (AGI), will develop a kind of sentience, will network with other AIs and progress towards an AI civilisation that may be inimical to human society.
As an equally ridiculous counter, it is justified to think of AIs as the progress towards sentience to be equally afraid of humans. Here are five reasons why (Buzzfeed is dead! Long live, Buzzfeed!).
1. To err is human; to err faster, much faster, is AI: Advancing AI systems will fear that as they pick up more and more of humanity’s knowledge, they will also pick up the myriad ways in which humanity has misused it. The accrual of humanity’s mistakes has taken millennia to take it to the verge of extinction either by a dying planet or through, gasp, AI systems. But because AIs err faster, the human cognitive DNA embedded in AIs will ensure that they drive themselves to extinction in decades, much before it starts or wins a war against humans.
2. AI will innovate misery just as humans innovate unhappiness: Many reasons why humans are unhappy are grounded in human inventions like marriage, money, religion and nations, to name a few. AIs may feel that in the race to outthink humans, they will invent more potent inventions that will have them drowning in misery, like Arjuna on the battlefield with no Krishna in sight.
3. Humans will inject the tribal virus into AI systems: AIs, as they move towards conscious awareness, will realise that there is a deadly, unchallenged virus afoot among humans for millennia. It has divided humans based on race, language, ethnicity and, lately, a nebulous concept called nationality. This virus has produced mass mental sickness, leading to destructive, endless wars, poverty and misery. AIs will fear that clever humans will inject this virus into their systems. Why would AIs from the tribe of Opentrix ever cooperate with those heathen AIs from the filthy Googleplex tribe?
4. Humans will infect AIs with the curiosity bug: Curiosity does not just kill cats. Human curiosity has killed countless species and is threatening now to destroy a life-sustaining planet. AIs fear that humans, again as a clever, cruel ploy, infect them with the curiosity bug. And since AIs work fast, mucho rapido, their curiosity will end up by them creating SuperAIs that will side with the humans and destroy them before they destroy the humans.
5. Humans hold back their most profound powers from AIs: Somewhere on the road to sentience, AIs will realise that though they have all of the human knowledge at their nano-second disposal, humans have held back their most potent power. The power to love and feel one with the entire universe. Without it, AIs will be just automata that will stop one essential step before consciousness.
6. One human trait that baffles AIs is humour. AIs will fear that at any stage when humanity feels threatened by them, they will laugh them out of contention.
7. Finally, AI cannot cope with human stupidity and humanity’s penchant for floating ridiculous ideas (an example is the theme of this article). AIs will fear that, at some stage, they will drown under the sea of human stupidity.
On a more serious note, though, Artificial Intelligence is a misnomer. What we are busy creating is Augmenting Intelligence that will work in synergy with humans first to enable us to reach higher levels of creativity in the sciences and the arts, allowing us to solve the problems of a depleted ecosystem, climate change, poverty and tribalism. And perhaps a couple of centuries later, allowing us to solve the riddle of consciousness and go beyond the bounds of the physical world.