Here are three tests for checking if a machine has conscious intelligence, in no particular order.
The Deep Thought test of conscious intelligence
In Douglas Adams’ The Hitchhiker’s Guide to the Galaxy, we learn that the Earth is just a giant computer made to produce the Ultimate Question (the answer was previously computed by other computer named Deep Thought). However, five minutes before the conclusion of the 10-million-year program, the Earth is destroyed by Vogons.
The following test is similar and goes like this:
- Create bunch of A.I. (Artificial Intelligence), say 10,000 of them.
- Put them in virtual reality (they will end up putting us in virtual reality anyway, so don’t feel bad for it).
- Wait, say… 3.000 years.
- Check their published philosophical virtual-works.
- Has anyone mentioned “qualia”? If yes, they have conscious intelligence.
Vogon fleet can destroy the Earth in meantime
Voight-Kampff test of conscious intelligence
Remember the movie Blade Runner ?
In that movie Rick Deckard, played by
Indiana Jones Harrison Ford, was hunting down few remaining replicants. Replicants in the movie are genetically-manufactured beings created by humans to do dangerous work at the “off-world” colonies. They are created to live only 4 years, so that they can’t develop their own “emotional responses”. But after a mutiny, special police units (Blade Runners) are sent to kill all the replicants. As they are same with humans in all respect except their emotions, Blade Runners use special Voight-Kampff test. In the test bodily responses such as heart rate, eye movement etc. are observed while asking “carefully worded questions and statements”.
Good just for replicants. I’m mentioning it mostly because I found that On The Edge Of Blade Runner (2000) documentary about the movie is available online. (As part of the Free Movies Fallen out of Copyright site…
Turing Test of conscious intelligence
Without intention to kill anyone, Turing proposed a test, which is today often mentioned in the context of the question – how can we know if machine has consciousness. In this test the machine(say a computer) which we are testing for consciousness and one (genuine) human are put in separate rooms, Then a Judge who communicates with both through text-only medium (e.g. Internet chat) asks them questions, trying to determine which is the human and which is the machine. In this test there can’t be “carefully worded questions and statements” which Judge could ask, as that would allow programmers to “carefully” program the replies. So, unlike the Blade Runners, the Judge will need to improvise and think of new smart questions as the test goes.
The Judge is human – once the humans are gone, A.I. won’t know if they have conscious intelligence or not.