Artificial Super Intelligence is the concept of an Artificial Intelligence that is more intelligent than humans. In an event coined The Singularity – not only will human intelligence be surpassed, but Artificial Super Intelligence will be capable of redesigning itself, accelerating it’s Intelligence to millions of orders of magnitude higher than ourselves.
It’s not a new concept, the idea of super intelligent machines has been part of Science Fiction for decades. The plot twist, that an Artifical Super Intelligence would through rational throught come to the conclusion that humans need to be eradicated is commonplace in fiction – and it has made its way into reality as well. No small number of people have warned against the claimed inevitable construction of an ASI and mankinds doom if we fail to control this technology.
There is however a huge flaw in the arguments. In short, there are some assumptions being made that should be disclosed.
The argument for Artificial Super Intelligence is simple. Looking at the current state of Artificial Intelligence, compare against the statistics regarding the human brain and then look at the exponential pace of technological advancement. With these three datasets, we can extrapolate a point in the future where this will occur. Quite simple, non?
However, todays Artificial Intelligence cannot do everything a human mind can. Today’s AI is not a scaled down version of the properties exhibited by the mind. Introspection, sense of self and consciousness is all essential properties in human intelligence – but not even close to be replicated by a machine. In fact, presuming that these qualities will emerge as we scale the complexity of current Artificial Intelligence implementations means that we make two big assumptions – we assume Materialsm and Determinism.
Materialism in philosophy is a monism which hold that all phenomena, including mental phenomena and consciousness, are results of material interactions. By creating an extrapolation from current AI to future Artificial Super Intelligence – logically, Materialism is assumed for mental phenomena.
Determinism in philosophy is a position that for every event there exist conditions that could cause no other event – since current AI is deterministic with the same model, extrapolating a line from current AI to future Artificial Super Intelligence also assumes that ASI will be deterministic.
By doing the extrapolating from AI to ASI we assuming no unknown forces or phenomena makes up human Intelligence – or at best, if such unknown forces or phenomena exists, they will also be replicable and able to be made into a machine by humans.
Keep in mind that both Materialism and Determinism is unresolved philosophical questions. Some philosophers even claim that if both where true, free will can’t exist and humans are not responsible for their own actions. Needless to say, we are entering a domain of existential questions. To assume the answers to unresolved existential questions – there is only one describing word: belief.
To take a position on Materialism and Determinism requires belief – it requires faith. The only places where you will find claims to have “solved” these questions is in fact in the world’s many religions. By believing in ASI and The Singularity, you are also believing in Materialism and Determinism. I’m not advocating that people should not have belief though, I’m merely pointing out that whenever one is basing logical argumentation on belief – one should disclose those beliefs.
A belief in an answer to Determinism and Materialism is not the only thing ASI have in common with religion. The whole narrative, humans creating an offspring better than themselves that will ultimately bring on their demise – shares common traits with many religious narratives. There is a birth, there is a consequence for playing gods, there is the apocalyptic ending. In short, Artificial Super Intelligence and the event called The Singularity leading to the demise of humans is starting to look a lot like a modern day religion.
The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.