Date: Sat 8 Oct 1988 15:06-EDT From: AIList Moderator Nick Papadakis Reply-To: AIList@AI.AI.MIT.EDU Us-Mail: MIT LCS, 545 Tech Square, Rm# NE43-504, Cambridge MA 02139 Phone: (617) 253-6524 Subject: AIList Digest V8 #98 To: AIList@AI.AI.MIT.EDU Status: RO AIList Digest Sunday, 9 Oct 1988 Volume 8 : Issue 98 Machine Consciousness (6 messages) ---------------------------------------------------------------------- Date: 5 Oct 88 17:47:57 GMT From: mailrus!uflorida!usfvax2!mician@ames.arpa (Rudy Mician) Subject: Intelligence / Consciousness Test for Machines (Neural-Nets)??? I have a question that I know has been addressed in the past (and undoubtedly continues to be addressed): When can a machine be considered a conscious entity? For instance, if a massive neural-net were to start from a stochastic state and learn to interact with its environment in the same way that people do (interact not think), how could one tell that such a machine thinks or exists (in the same context as Descarte's "COGITO ERGO SUM"/"DUBITO ERGO SUM" argument- that is, how could one tell whether or not an "I" exists for the machine? Furthermore, would such a machine have to be "creative"? And if so, how would we measure the machine's creativity? I suspect that the Turing Test is no longer an adequate means of judging whether or not a machine is intelligent. If anyone has any ideas, comments, or insights into the above questions or any questions that might be raised by them, please don't hesitate to reply. Thanks for any help, Rudy -- Rudy Mician mician@usfvax2.usf.edu Usenet: ...!{ihnp4, cbatt}!codas!usfvax2!mician ------------------------------ Date: 5 Oct 88 20:24:56 GMT From: bwk@mitre-bedford.arpa (Barry W. Kort) Subject: Re: Intelligence / Consciousness Test for Machines (Neural-Nets)??? In article <1141@usfvax2.EDU> mician@usfvax2.usf.edu.UUCP, (Rudy Mician) asks: >When can a machine be considered a conscious entity? Consciousness is not a binary phenomenon. There are degrees of consciousness. So the transition from non-conscious to conscious is a fuzzy, gradual transition. A normal person who is asleep is usually regarded as unconscious, as is a person in a coma. An alert Dalmation may be considered conscious. It might be more instructive to catalog the stages that lead to higher levels of consciousness. I like to start with sentience, which I define as the ability of a system to sense its environment and to construct an internal map, model, or representation of that environment. Awareness may then be defined as the ability of a sentient system to monitor an evolving state of affairs. Self-awareness may, in turn, be defined as the capacity of a sentient system to monitor itself. As an aware being expands its powers of observation, it achieves progressively higher degrees of consciousness. Julian Jaynes has suggested that the bicameral mind gives rise to human consciousness. By linking two semi-autonomous hemispheres through the corpus callosum, it is possible for one hemisphere to act as observer and coach for the other. In other words, consciousness requires a feedback loop. Group consciousness arises when independent individuals engage in mutual mirroring and monitoring. From Narcissus to Lewis Carroll, the looking glass has served as the metaphor for consciousness raising. --Barry Kort ------------------------------ Date: 6 Oct 88 12:06:26 GMT From: uhccux!lee@humu.nosc.mil (Greg Lee) Subject: Re: Intelligence / Consciousness Test for Machines (Neural-Nets)??? >From article <1141@usfvax2.EDU>, by mician@usfvax2.EDU (Rudy Mician): " ... " When can a machine be considered a conscious entity? Always. It's a matter of respect and empathy on your part. All the machines I use are conscious. Or never, maybe, if you take 'conscious' seriously enough to entertain the possibility that you yourself are not conscious except sporadically. Whatever one may think of his overall thesis, Julian Jaynes (The Origin of Consciousness and the Bicameral Mind) is very persuasive when he argues that consciousness is not required for use of human language or every-day human activities. Greg, lee@uhccux.uhcc.hawaii.edu ------------------------------ Date: 6 Oct 88 15:06:55 GMT From: tank!uxc!ksuvax1!cseg!cdc@oddjob.uchicago.edu (C. David Covington) Subject: Re: Intelligence / Consciousness Test for Machines (Neural-Nets)??? In article <1141@usfvax2.EDU>, mician@usfvax2.EDU (Rudy Mician) writes: > > I have a question that I know has been addressed in the past (and undoubtedly > continues to be addressed): > > When can a machine be considered a conscious entity? > . . . > > I suspect that the Turing Test is no longer an adequate means of judging > whether or not a machine is intelligent. > Regarding intelligent machines, to the naive it's totally magic, to the wizard it's clever programming and a masterful illusion at best. To ascribe consciousness to a machine is a personal matter. If I cannot tell the difference between a 'conscious' human and a skillful emulation of the same, then I am perfectly justified in *modeling* the machine as human. It's not so much a question of what *is* as a question of what *appears* to be. The same machine might be rightfully deemed conscious by one but not by another. I must expose my world view as predominantly Christian at this point. My belief in a Supreme Being places my view of man above all other animals and therefore above any emulation of man by machine. I say this not so much to convert the masses to my point of view but to clarify that there are people that think this way and this allows no place for conscious machines. So to readdress the original question, the Turing test is certainly still valid from my understanding that it is a matter of how accurately you can mimic human behaviors. Between the lines you are making the assumption that man and machine are the same in essence. To this I object by faith. The question cannot be properly addressed without first dealing with world views on man. David Covington Assistant Professor Electrical Engineering University of Arkansas (501)575-6583 ------------------------------ Date: 6 Oct 88 23:58:51 GMT From: esosun!jackson@seismo.css.gov (Jerry Jackson) Subject: Re: Intelligence / Consciousness Test for Machines (Neural-Nets)??? In article <40680@linus.UUCP> bwk@mitre-bedford.ARPA (Barry W. Kort) writes: In article <1141@usfvax2.EDU> mician@usfvax2.usf.edu.UUCP, (Rudy Mician) asks: > >When can a machine be considered a conscious entity? > A normal person who is asleep is usually regarded as unconscious, > as is a person in a coma. An alert Dalmation may be considered > conscious. A person who is in a coma is unconscious because he is incapable of experiencing the outside world. Consciousness is a *subjective* phenomenon. It is truly not even possible to determine if your neighbor is conscious. If a person felt no pain and experienced no colors, sounds, thoughts, emotions, or tactile sensations he could be considered unconscious. Note that we would be unable to determine this. He could behave in exactly the same way while being completely inert/dead inside. Machines that are obviously unconscious such as feedback-controlled battleship guns and thermostats respond to their environments but, I would hardly call them conscious. It is hard to imagine what one would have to do to make a computer conscious, but it does seem that it would involve more than adding a few rules. --Jerry Jackson ------------------------------ Date: 7 Oct 88 20:09:54 GMT From: hubcap!ncrcae!gollum!rolandi@gatech.edu (mail) Subject: Re: Intelligence / Consciousness Test for Machines (Neural-Nets)??? Adding to Barry Kort's...... >Consciousness is not a binary phenomenon. There are degrees of >consciousness. So the transition from non-conscious to conscious >is a fuzzy, gradual transition. When a person is awake and responds in a predictable manner, he is said to be conscious. >Awareness may then be defined as the ability of a >sentient system to monitor an evolving state of affairs. When a person is known to know some given thing, he is said to be aware. >Self-awareness may, in turn, be defined as the capacity of a sentient >system to monitor itself. When a person can label his own behavior in ways that are consistent with the labels of those who observe him, he is said to be self-aware. Walter Rolandi rolandi@ncrcae.Columbia.NCR.COM NCR Advanced Systems Development, Columbia, SC ------------------------------ End of AIList Digest ********************