My intuition is that 1) self-awareness (modeled through epistemic modal logics); 2) goal reasoning (e.g. https://www.aaai.org/ocs/index.php/AAAI/AAAI16/paper/view/12292) and 3) consciousness (e.g. Integrated Information Theory) are the necessary and sufficient features for agent to be truly autonomous. Especially - consciousness is the key feature of autonomy. So - can truly autonomous (human-level autonomy) agent exist without consciousness?

Of course, I am thinking about autonomy not ultimate freedom. E.g. every agent can be autonomous but it will be bounded by 1) scarcity of resources; 2) ultimate goals and values which will be inscribed into agent by the humans.

  • 1
    $\begingroup$ This sounds like a question of philosophy, not a technical question about computer science. At least it does not seem like a well-posed question without a definition of "truly autonomous agent" and "consciousness". $\endgroup$
    – D.W.
    Commented May 28, 2018 at 21:41
  • $\begingroup$ Integrated Information Theory defines quantitative measures of consciousness. The only question is - is this consciousness the essential component of autonomous agent? Human level AI is considered in AGI (published in LNCS0 conferences and partially also BICA conferences. $\endgroup$
    – TomR
    Commented May 28, 2018 at 22:24
  • $\begingroup$ Yeah this is more a question of philosophy than computer science. $\endgroup$
    – Lance
    Commented May 29, 2018 at 1:15
  • 1
    $\begingroup$ If you want to ask about consciousness in the sense defined by "Integrated Information Theory", please edit the question to (a) state that, and (b) provide that definition of consciousness. Then, provide a similar definition of "truly autonomous". Otherwise, there are too many things that "consciousness" could possibly mean; different people probably interpret the same word differently, so it's hard to identify a concrete answerable question if you don't define your terms. $\endgroup$
    – D.W.
    Commented May 29, 2018 at 20:08
  • $\begingroup$ Interesting question,though it requires some revision a little bit.,However,I kindly request the OP to bring such questions @artificial intelligence community,for effective feedback. $\endgroup$
    – quintumnia
    Commented Jun 2, 2018 at 17:56

1 Answer 1


Here is a glimpse.

Concerning agents you're trying to elaborate here,I would like to give you some insights of these following key aspects,which might help in your future research nor knowledge;

  • Consciousness Here an agent is capable of understanding its own existence nor selfawareness.

  • Intelligence An agent is capable solving problem based reasoning with no help from third-party knowledge.

  • Autonomous An agent has the ability to solve problems basing on the third-party knowledge for instance,self driving car does its own problem solving basing on the existing knowledge it discovers from the environment.

Therefore,for any agent to be autonomous,it doesn't mean its selfaware or conscious,otherwise. For your information,do further reading on the above three terminologies.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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