Week 7 Readings

Nick Bostrom, Superintelligence; Paths, Dangers, Strategies &
Portrayals and perceptions of AI and why they matter, The Royal Society

Bostrom starts off his book with The Unfinished Fable of the Sparrows, in which he attempts to introduce how the pursuit of a more intelligent and capable being- in the story’s case, the owl. The pursuit to assimilate a foreign intelligent entity to help lift the sparrow’s everyday burdens. In the story, the majority of the sparrows fell into the romantiscized Idea of assimilating the owl into the flock without much thought and proceeded to look for an owl egg- but only one, Scronkfinkle was skeptical of the domestication of an owl and demanded for more thought to be put into the implementation before bringing the owl into their midst. 

In the Royal Society paper the workshop conducted has determined that the public framework of the AI narrative has spanned between fear and hope. The extreme fears represent Scronkfinkle apprehension around the assimilation of the foreign being, similar to the public’s opinion on the fear and uncertainty that AI entails. Namely, “the loss of humanity; making humans obsolete; alienating people from each other; and enslaving or destroying humans.” Bostrom believes that with the advancement of machines at the current rate the imminent ability for machines to match or even surpass humans intelligence is just around the corner. However, without an assessment for risk to safety concerns or ethical doubts in relation to the creation of artificial minds or even- potential computer overlords. I posit that to assume the emergence of computer overlords is to assume that AI is capable of ‘thinking’ for itself with the tools and data we have provided thereby allowing it to  transcend beyond human intelligence. The dangers arises from a machine’s or AI’s lack of ability or disregard to nuances and empathise according to our moral ethics. The lack of this ‘human’ element is what people are afraid of because it only allows for the machine the ability to make ruthless decisions based on its painfully pragmatic approach to problems. 

“A mind that is limited to reason and analysis is incapable of perceiving what is truth” – J. Krishnamurti 

Our perception of truth includes compassion and a machine is unlikely to have this capability, even if it does, it would be corrupt by humans eventually. No human is a perfect being, and AI will learn to mimic this ‘imperfect’ behaviour which translates to the data it receives or even the foundations of morals we chose to implement within the system which are littered with human biases. This ‘human’ element has been mimicked in the past using ELIZA, a computer program that emulates a Rogerian psychotherapist Therapist. Yet unsuccessfully, below is an example of how the program works and it displays a rather indifferent facade and feels like talking to yourself. When Ke Jie, the world champion of Go was defeated by AlphaGo Ke Jie claimed that AlphaGo “Sees the whole universe- while we only see a pond infant of us, so I’ll leave it with the task to explore the universe while I, myself prefer the fishing pond” we understand that AI’s ability to surpass our intelligence has led us to believe that AI poses God-like attributes when in reality its merely nothing more than simple and narrow algorithms there is no need for AI to understand abstract concepts or clever strategy to the game. 

On the other hand of the spectrum, the enticing idea and hope that AI could be the solution to “ageing and disease so that humans might lead vastly longer lives; freeing humans from the burden of work; gratifying a wide range of desires, from entertainment to companionship; and contributing to powerful new means of defence and security.” In a Ted talk video named ‘how AI can save our humanity’ by Kai-Fu Lee explained an interesting prospect for the advancement to AI. Jobs that are repetitive, routine based and focus on optimisation will eventually be replaced by AI within the next decade. However, complex and creative jobs will not be dominated by AI because it is only capable of optimising and not creating and when we allow for AI to take over routine based jobs we are allowing ourselves the opportunity to do meaningful work in the field of compassion. To see a possibility to work with AI and coexist in the world, instead of impending doom- AI gives us freedom.