Jump to content


Photo

Is it a person or is it not?

  • Please log in to reply
112 replies to this topic

#1 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 23 July 2014 - 03:36 PM

I am participating in an on-line asynchronous e-learning program. A discussion is ensuing about the superiority of this type of learning/knowledge-acquistion to face-to-face dialogical synchronous learning/knowledge acquisition.

Relevant to this discussion is the question of if and when verbalizing-machines with advanced-software will reach a point, in such an on-line environment, one will no longer be able to tell if one is speaking with another person or such a machine.

I am interested in your opinions about if and when such a "moment" will be reached; when the ability to distinguish will no longer be possible.

Otto

#2 Pess

Pess

    (Title)

  • *****
  • Posts: 2881
  • Joined: 12 Sep 2007
  • Loc: Toledo, Ohio

Posted 23 July 2014 - 04:09 PM

You are talking about the Turing Test.

In basic conversations (especially between interviwer vs interviewee) it has already been shown that it is very difficult to differentiate between advanced AI's and humans in a conversation.

In fact, today many news stories are 'written' by AI's that cull news bits from various stories and write original articles that are almost impossible to differentiate from human written ones.

Pesse (When an AI writes an original novel get back to me) Mist

#3 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 23 July 2014 - 04:12 PM

I strongly suspect that I have had phone conversations with such-machines, for which the interactions were so well scripted in the software that I entered into the conversation for a few moments before realizing it was a machine.

I'm wondering, more specifically, when the day will come when on-line asynchronous facilitators of e-learning will be such-machines and the "students/participants" will not be able to tell if it is such a machine.

#4 Pess

Pess

    (Title)

  • *****
  • Posts: 2881
  • Joined: 12 Sep 2007
  • Loc: Toledo, Ohio

Posted 23 July 2014 - 04:35 PM

I dunno. I imagine they could come up with a pretty good AI professor in a math class.

But it is all cost/benefit ratio. What is cheaper? Underpay some poor teacher to reach 10K kids online or pay an IT specialist to maintain the program on the server?

Pesse (AI teachers might be awhile) Mist

#5 GJJim

GJJim

    Surveyor 1

  • -----
  • Posts: 1681
  • Joined: 09 Sep 2006
  • Loc: Western CO

Posted 23 July 2014 - 04:53 PM

After suffering through math classes taught by a professor that stuttered, a professor with an Asian accent and poor English skills, and an Austrian fellow that assumed everyone could read the notes he handed out from a French textbook, a 'bot might be an improvement! :grin:

#6 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 23 July 2014 - 04:56 PM

I think it is a terrifically interesting ethical and political argument about whether or not such-machines should be used instead of human teachers.

But, first, (I hope you don't mind the insistence) I would like to hear more and from others about if and when such-machines will become indistinguishable from a human (on-line asynchronous e-learning) presence.

Thanks gals and guys.

Otto

#7 Elric82

Elric82

    Vostok 1

  • -----
  • Posts: 152
  • Joined: 25 Sep 2011
  • Loc: Richmond, Kentucky

Posted 23 July 2014 - 05:09 PM

Nice to see a fellow central Kentuckian on here!

#8 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 23 July 2014 - 05:53 PM

Pleased to make your acquaintance.

How much damage was done downtown in the fire a few weeks back?

#9 GregLee1

GregLee1

    Viking 1

  • -----
  • Posts: 669
  • Joined: 21 Jul 2013
  • Loc: Waimanalo, HI

Posted 23 July 2014 - 06:35 PM

It's very difficult, without imposing constraints on the communication that conceal the details which a human could use to distinguish another human from a device. (Recall that Eliza had to use a Teletype.) There are two big problems:
  • (1) We have evolved special abilities to tell one of our tribe from a stranger -- witness how good we are at detecting foreign accents. We're very good at it. It's a survival characteristic.
  • (2) We don't understand very well the subtleties of speech that people can use to determine characteristics of someone talking. If we don't understand it, this makes it difficult to program.
I'll predict we will not achieve it in my lifetime.

#10 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 23 July 2014 - 08:29 PM

I appreciate your thought-filled response, Greg.

Other folk?

Otto

#11 petrus45

petrus45

    Mariner 2

  • *****
  • Posts: 240
  • Joined: 31 Jan 2011
  • Loc: SW Ohio / N Ky.

Posted 23 July 2014 - 09:08 PM

Here's how you'll know what to do. Allan Sherman's "Automation" at 2:45.
http://www.youtube.c...0894EAEE32551D7

#12 ColoHank

ColoHank

    Vanguard

  • *****
  • Posts: 2330
  • Joined: 07 Jun 2007
  • Loc: western Colorado

Posted 23 July 2014 - 09:42 PM

Call me old-fashioned, but I like human-to-human interactions, and I also like the lasting bonds I formed with many of my teachers and professors. I can't imagine that any machine, no matter how carefully and realistically it's programmed, could match any of them.

#13 llanitedave

llanitedave

    Humble Megalomaniac

  • *****
  • Posts: 22580
  • Joined: 25 Sep 2005
  • Loc: Amargosa Valley, NV, USA

Posted 23 July 2014 - 10:59 PM

I prefer a good human to the best robot. But I prefer a mediocre robot to the typical self-serving bureaucrat/politician.

#14 darknesss

darknesss

    Ranger 4

  • -----
  • Posts: 303
  • Joined: 05 Sep 2013

Posted 24 July 2014 - 01:06 AM

I am participating in an on-line asynchronous e-learning program. A discussion is ensuing about the superiority of this type of learning/knowledge-acquistion to face-to-face dialogical synchronous learning/knowledge acquisition.

Relevant to this discussion is the question of if and when verbalizing-machines with advanced-software will reach a point, in such an on-line environment, one will no longer be able to tell if one is speaking with another person or such a machine.

I am interested in your opinions about if and when such a "moment" will be reached; when the ability to distinguish will no longer be possible.

Otto

You sound like a robot......... :)

#15 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 24 July 2014 - 02:34 PM

With Hank and Dave, I prefer the dialogical interaction that occurs when one has an informed educator working with a group of persons engaged in a courteous, but not the less vigorous, give and take.

So far, I sense that the opinion here is that the automated/programmed/software/machine for online asynchronous instruction will not soon reach a level of sophistication in which most persons could not distinguish it from a human teacher.

Or, are there some of you/us here who think the day is coming where in non-face-to-face asynchronous on-line education, the day is coming where a program/software/machine can replace the teacher to such a degree few could tell it was not a person?

Otto

#16 GregLee1

GregLee1

    Viking 1

  • -----
  • Posts: 669
  • Joined: 21 Jul 2013
  • Loc: Waimanalo, HI

Posted 24 July 2014 - 02:59 PM

With Hank and Dave, I prefer the dialogical interaction that occurs when one has an informed educator working with a group of persons engaged in a courteous, but not the less vigorous, give and take.

Why must the educator be informed? This idea of education seems to view education as a process of knowledge flowing down from teacher to student. Think of the most important things you've ever learned. Were they things that a teacher knew and conveyed to you?

#17 ColoHank

ColoHank

    Vanguard

  • *****
  • Posts: 2330
  • Joined: 07 Jun 2007
  • Loc: western Colorado

Posted 24 July 2014 - 03:44 PM

Or, are there some of you/us here who think the day is coming where in non-face-to-face asynchronous on-line education, the day is coming where a program/software/machine can replace the teacher to such a degree few could tell it was not a person?



I suppose that, given enough time and persistence, it might be possible to program an on-line terminal to interact so convincingly with humans that a distant student might not be able to tell the difference between machine and man. But to what avail?

Do we really need to replace teachers with machines? I know it's cheaper to plug a great many students into an on-line course than it is to have one instructor interact with just a few at a time in the confines of a physical space. But will that multitude of on-line students receive better educations, or educations that are lacking? If their educations are better, what use will all of them be able to make of them? (There sure won't be much demand for teachers, will there?)

Who will profit, and who won't? Will the savings be utilized for the greater good, or will they just line the pockets of a few?

Is it really desirable to replace ever more humans with machines? If it is, then what will those displaced humans do to remain productive and useful? How will they earn their livings in a world that's increasingly automated?

Should we do things only because they're possible, or should we be motivated instead by genuine need?

So many questions...

So few answers.

#18 Pess

Pess

    (Title)

  • *****
  • Posts: 2881
  • Joined: 12 Sep 2007
  • Loc: Toledo, Ohio

Posted 25 July 2014 - 11:32 AM

With Hank and Dave, I prefer the dialogical interaction that occurs when one has an informed educator working with a group of persons engaged in a courteous, but not the less vigorous, give and take.

So far, I sense that the opinion here is that the automated/programmed/software/machine for online asynchronous instruction will not soon reach a level of sophistication in which most persons could not distinguish it from a human teacher.

Or, are there some of you/us here who think the day is coming where in non-face-to-face asynchronous on-line education, the day is coming where a program/software/machine can replace the teacher to such a degree few could tell it was not a person?

Otto


Since no AI of sufficient complexity exits yet, how do any of you know what you'd prefer?

Sure, I'd like to talk with a Hot Chick over my Apple IIe right now, but the Replicant Pris Stratton might have an edge...

In any event, does anyone want to invest in my new start-up company?

Pesse (I think I'll name it the Tyrell Corporation) Mist

#19 Qwickdraw

Qwickdraw

    Surveyor 1

  • *****
  • Posts: 1729
  • Joined: 03 Mar 2012
  • Loc: Ann Arbor, MI

Posted 25 July 2014 - 11:43 AM

"Is it a person or is it not?"

I often have to ask this question when I watch the evening news.

#20 FirstSight

FirstSight

    Duke of Deneb

  • *****
  • Moderators
  • Posts: 9916
  • Joined: 26 Dec 2005
  • Loc: Raleigh, NC

Posted 28 July 2014 - 10:46 PM

Relevant to this discussion is the question of if and when verbalizing-machines with advanced-software will reach a point, in such an on-line environment, one will no longer be able to tell if one is speaking with another person or such a machine.


One of the key factors toward helping machines potentially succeed in achieving this goal in an online environment is the extent to which extraneous contexts beyond the task-at-hand can be avoided in its interaction with a human participant. An example will help clarify what I mean: many of the online (human) customer service reps which some domestic American airlines now employ are are actually housewives working online out of their kitchens - something I learned when I ran into a longtime employee of my local YMCA health club, whom I hadn't seen in awhile, on a visit to a nearby grocery store. She was now working for one of the airlines for their rewards club, working from home. Other than perhaps the absence of subdued background noises of other customer reps, you'd never be able to tell she was sitting in her kitchen in shorts and a T-shirt physically alone, rather than working somewhere in a cubicle or bank of reps. The relevant point is: in your normal on-task interactions with her, there is a near complete absence of any context that would contradict your natural assumption that she is working in some sort of collegial office environment - she has immediate online voice or IM communication with supervisors to help solve any problems helping customers in the same sort of real-time frame while the customer is still on the line as she would at a collegial call center. The other key thing that helps convincingly "sell" the impression (or assumption) that she's based in a professional (live) collegial call center rather than physically alone in her kitchen is the unlikelihood, in their interactions with her, of customers departing from the strict range of airline rewards program issues they contacted her about.

Likewise, a machine is much more likely to be successful convincingly interacting with a human participant in an online course if the possibilities for the human to take discussions beyond a relatively narrow context are minimized. For example, it's much more likely a machine can be programmed to convincingly act to simulate a human calculus tutor to a confused student than to successfully simulate a human philosophy tutor. It's going to be a long time before someone comes up with a philsosophy-course interactive tutor that will convincingly fool Otto P. that he's dealing with a human rather than a machine.

#21 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 29 July 2014 - 07:47 AM

Thanks Chris. Your's is the type of analysis that helps me get to the root of the issues to be considered.

Otto

#22 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1028
  • Joined: 18 May 2005
  • Loc: Kansas, USA

Posted 29 July 2014 - 02:26 PM

"Is it a person or not?" The answer is "not." A person, from the biological perspective is a member of the species Homo sapiens, genealogically linked to a set of parental Homo sapiens(and also genealogically linked, albeit remotely, to oak trees). It may be an individual entity, and perhaps one as smart as a person. And, you might even think you are speaking to a person when you speak to it. Oh, by the way, corporations are not persons from the biological perspective either. They may be individuals (under the law), but they certainly are not people.

Ed

#23 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 29 July 2014 - 02:38 PM

The title "Is it a person or is it not" was a tongue-in-cheek reference to an ancient advertisement regarding the cassette music/audio tape called Memorex; i.e. "Is it live or is it Memorex" for the purpose of introducing the opening question found in the opening post. The title was not meant to be taken literally.

Otto

#24 scopethis

scopethis

    Cosmos

  • *****
  • Posts: 8641
  • Joined: 30 May 2008
  • Loc: Kingman, Ks

Posted 29 July 2014 - 07:29 PM

as an ancient one, I do recall those ads; I also remember the debate about plastic pickles on "fast food" burgers...

#25 Otto Piechowski

Otto Piechowski

    Apollo

  • -----
  • Posts: 1463
  • Joined: 20 Sep 2005
  • Loc: Lexington, KY

Posted 29 July 2014 - 09:20 PM

I don't remember the plastic pickles on fast food burgers. What was that one?

Otto






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics