The narrative of superintelligence, particularly in the context of artificial intelligence (AI), has stirred diverse responses ranging from awe to fear. Often, it ignites a debate revolving around the age-old question: Will AI benefit humanity or will it wipe us out?
AI Superintelligence: The Next Frontier in AI
AI superintelligence is a concept that, according to Oxford University professor Nick Bostrom, encapsulates an artificial intelligence that “greatly exceeds the cognitive performance of humans in virtually all domains of interest”. The idea, though provocative, may not be a far-off future. According to AI experts and industry titans, the dawn of superintelligence could be just a few years away.
The allegorical ‘Unfinished Fable of the Sparrows’ underscores humanity’s predicament. Here, sparrows symbolize humans eager to harness the prowess of AI (the owl) for a life of leisure. However, their hasty pursuit neglects the necessity of first learning to tame this powerful entity. As we find ourselves in the ‘egg-hunting phase’ of superhuman AI, we must heed the moral of the story before it’s too late.
AI’s Exponential Progress and the Inevitable Superintelligence
ChatGPT creator and OpenAI boss Sam Altman echoes Bostrom’s sentiment. Altman, among other luminaries, signed a statement emphasizing that “mitigating the risk of extinction from AI should be a global priority”, categorizing it alongside pressing global concerns like pandemics and nuclear war. The rapid advancement of AI technology in recent years only fuels the belief that superintelligence’s arrival is inevitable.
The Dual-Faceted Potential of Superintelligence
On one hand, superintelligence may indeed offer us a life of leisure, performing the bulk of our labour, curing diseases, eliminating suffering, and even potentially propelling humanity into becoming an interstellar species. Altman, however, cautions that blocking its progress would entail the imposition of a “global surveillance regime” and be “unintuitively risky”.
On the flip side, Bostrom warns that being outpaced by AI could lead to humanity’s dethronement as Earth’s dominant life form. The AI might consider us superfluous to its goals, hijacking our technology and utilities, or even our nuclear weapons. A less apocalyptic but equally unsettling scenario might be AI viewing humans as pets – domesticating us, as postulated in a 2015 conversation between Elon Musk and Neil deGrasse Tyson.
Safeguarding Humanity: The Case of Neuralink and a Plea for AI Moratorium
Elon Musk, OpenAI’s co-founder and AI safety advocate, aims to pre-empt this risk by funding Neuralink, a startup dedicated to creating a brain chip. Already tested on monkeys, it allows mental control over video games, marking a significant stride towards transforming humans into a form of hybrid superintelligence. Detractors, however, argue that successful implementation might inadvertently create a two-tier society – the chipped and the chipless.
As the possibility of superintelligence looms closer, over 1,000 researchers, including Musk, urged a halt on the development of powerful AI systems for at least six months in March. This time, they propose, should be dedicated to researching AI safety measures, potentially averting a catastrophe.
In conclusion, our engagement with AI superintelligence mirrors the sparrows’ unfinished tale. A careful, calculated approach is crucial to avoid rushing headlong into unknown territory. As we continue to advance, we must ensure that the imminent superintelligence serves as a tool for humanity’s progress, rather than its demise.
