... that you would believe! As far as I know, we really don't have something akin to a Turing Test for sentience. A blinded test where a human can't differentiate between a human and a machine really isn't a test for sentience. I would expect something unexpected that can't be mined out of massive databases or chatbot training. I would be impressed by a statement concerning spirituality in first person or a fear of being unplugged. An exceptional insight into Pascal's Wager would be impressive. So to would be an explanation of our perception of the forward movement of time, which seems to have no basis in known physics. On the other hand, showing us a ToE could be just number crunching. There is also the question of whether we would perceive an AIE to be sane by human standards.

  • 2
    Sentient or sapient? A spider or a bat are sentient... – Philip Klöcking Mar 17 at 18:22
  • print("Hey I'm conscious in here, send in pr0n and LOLCats!") – user4894 Mar 17 at 19:10
  • Anything "could" be just number crunching, with enough resources any performance can be imitated. But performing a variety of tasks from an unspecified in advance list for a prolonged period of time, like average humans do, would be impressive enough in practical terms. – Conifold Mar 17 at 19:49
  • Yes, @PhilipKlöcking . We don't know what the minimum requirements for sentience are. Maybe paramecia are sentient. Maybe computer programs are already sentient. Sentience does not mean intelligence, it means the capability to have subjective experiences, i.e. qualia. One can perceive qualia without thinking about much of anything at all. – causative Mar 17 at 21:30

There are various versions of the Turing test, including modern extended versions about being able to make efficient abstractions. People miss Turing's real core point though, which is that we have to find a way to use evidence, rather than imagining there is a difference in essence (eg, a soul). That is the core of the Turing Test, and on that basis it will evolve as AGI does.

'Spirituality' is a woolly term, I'd substitute maybe addressing existential concerns. A theist is an atheist of every other religion, Pascal's wager is only if extremely parichial concern. Fear of pain & death are meat-robot concerns programmed by DNA replicating that has those priorities. Time, I'll give you - uniting time with quantum field theory is probably the biggest question in physics. Might it not be about what it does with such deeper knowledge - and it then be too late for us to pass judgement on a true AGIs sanity, and time for us to accept such a beings judgement of ours? Perhaps when it gently but firmly takes away nuclear weapons from us forever, which we even recognised as only for 'mutually assured destruction', the definition of madness.

Bostrom covers how unlike our own minds AGIs might be in Superintelligence: Paths, Dangers, Strategies, in particular relating to hidden suffering as a 'mind crime', that could result in a perverse instantiation and malignant failure mode.

We have a long way to go before we can really answer this question, and it's likely there will be a whole ecology of intelligences, just as in the border of what is called life there are RNA & viruses - yet, wherever a 'line' is we have pragmatic contextual definitions that depend on what we are doing.

I like the Neuralink approach of actively merging human and silicon-based intelligence. Based on intelligence as rooted in intersubjectivity, I would say humans and AGIs being able to share each others experiences will be very helpful. Just as we know dolphins are intelligent but we don't know how intelligent without being able to share a language, we will need the equivalent with AGI & Neuralink may help get it (frequency analysis suggested dolphins communicate with information density of 3 year old humans, but then it was discovered they use 3D sound holograms to communicate, disrupting our assumptions of using language like we do).

We need a whole set of questions about how minds work to get to an answer, and likely we will be wrong either by being too sceptical or not enough, when AGI emerges. The world will change though, of that be sure.

Some related discussions:

Intelligence vs capacity for intentionality/identity Free Will and Intelligence

Self-awareness What's the importance of self-awareness?

What is intelligence, AI vs AGI What is intelligence?

What meaning is According to the major theories of concepts, where do meanings come from?

Intersubjectivity Philosophy and the question 'When is a robot considered alive and thinking?'

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.