The term “Artificial General Intelligence” or AGI is the often-used yet rarely understood “goal” that contemporary AI researchers and companies claim to be working towards. Sadly, no one can agree on what it actually means.
Geoffrey Hinton, a computer scientist formerly employed by Google who is often referred to as “the godfather of AI”, says AGI is not a technical term but “a serious, though ill-defined, concept. I don’t think there is agreement on what the term means. I use it to mean AI that is at least as good as humans at nearly all of the cognitive things that humans do.”
Most definitions of AGI tend to point towards it being a type of Artificial Intelligence that is “as intelligent as human beings” at certain tasks. In reality, this widely-accepted definition is intentionally vague and nearly impossible to qualify, as science has not yet been able to agree on exactly what human intelligence is. There is also no consensus on which human tasks AGI will be capable of; some definitions stick to tasks of logic or problem-solving, while others couch the tasks in the language of Capitalism and “productivity”. So is Capitalism what makes us human? Were pre-Capitalist humans somehow not human? Are contemporary people who oppose Capitalism somehow inhuman?
One recent academic attempt to construct a definition of AGI compares it to the intelligence level of a “well-educated adult“. But by whose standard is someone “well-educated”? Within which culture/tradition/race/economic class is this “well-educated adult”? How old are they? What do they specifically know and not know? As should be apparent by this point, AGI is a concept that is very easily manipulated to adhere to any number of political and social ideologies, all of which are utterly subjective.
Several AI theorists predict that we will “know it when we see it”, and that, in all likelihood, AGI won’t just be intelligent, but “superintelligent” (i.e. all-knowing), akin to a god, and humans will somehow either be forced, or volunteer, to ‘merge’ with this new deity, forming a post-human race. Seriously, they believe this is inevitable. And right now the strategy is to just keep building larger and larger data centers until someone ends up creating this new technological god, who will then (obviously) teach us how to travel to other planets and colonise them with our superior technology (because of course, any god created by humans would immediately view colonialism as its only logical course of action).
This is not exaggeration. This is literally what most US-based AI company executives believe.
Watch AI sceptic Cory Doctorow’s rebuttal to the idea that AI will inevitably become conscious: