På AI-teknikens axlar : Om kunskapssociologin och stark artificiell intelligens
(2009) In Lund Dissertations in Sociology 87.- Abstract
- This dissertation is concerned with Sociology’s stance in the debate on Strong Artificial Intelligence, i.e. such AI that is able to shape new knowledge without human interference. There is a need for sociologists to realize the difference between two approaches to constructing AI systems: Symbolic AI (or Classic AI) and Distributed AI – DAI.
Sociological literature shows a largely critical attitude towards Symbolic AI, an attitude that is justified. The main theme of this dissertation is that DAI is not only compatible with Sociology’s approach to what is social, but also constitutes an apt model of how a social system functions. This is consolidated with help from Niklas Luhmann’s social systems theory and from Vygotsky-oriented... (More) - This dissertation is concerned with Sociology’s stance in the debate on Strong Artificial Intelligence, i.e. such AI that is able to shape new knowledge without human interference. There is a need for sociologists to realize the difference between two approaches to constructing AI systems: Symbolic AI (or Classic AI) and Distributed AI – DAI.
Sociological literature shows a largely critical attitude towards Symbolic AI, an attitude that is justified. The main theme of this dissertation is that DAI is not only compatible with Sociology’s approach to what is social, but also constitutes an apt model of how a social system functions. This is consolidated with help from Niklas Luhmann’s social systems theory and from Vygotsky-oriented education scientists who claim that processes leading to new knowledge are about expansion and not about problem solving. Problem solving only leads to elaborating existing knowledge. It is shown that human being has always used several types of artefacts and tools to produce culture and knowledge. This shows that humans are dependent on their environment and that knowledge is not only something that has to do with their brain, but also the rest of their bodies and their environments.
Further, Strong AI is not about robots thinking in the same way as humans, but about holistic social processes where the actor does not need to have a complex functionality. This provides a good opportunity to explain what sociologists call emergency, i.e. how social processes shape new knowledge independent of human actors.
The possibility of AI has to be evaluated in terms of human beings’ capacities to cope with reflexive processes. Luhmann teaches us that we have to see the difference between three different forms of reflexivity: self-reference, reflexivity and reflection. We contend that, in order to be able to shape new knowledge in expanding processes, there must be circumstances that allow reflection. Luhmann writes that knowledge-producing processes are dependent on capacities for reflection between the social systems and their environments. Strong AI has more capacity to handle these processes than humans have, while the strongest argument for strong DAI is that it can reach a far wider area than human beings’ brains can. This capacity for reflection is a better way of judging the capacity of strong AI, than questions about causal capacity and capacity for action. (Less) - Abstract (Swedish)
- Popular Abstract in Swedish
Avhandlingen diskuterar Sociologins position i debatten kring artificiell intelligens, d.v.s sådan AI som kan skapa kunskap på egen hand. Det påpekas att sociologin behöver göra en åtskillnad mellan två sätt att skapa AI system: Symbolisk AI (eller Klassisk AI) och Distribuerad AI – DAI.
Den sociologiska litteraturen visar en i stort kritisk attityd till symbolisk AI vilket är rätt. Avhandlingen handlar istället om att DAI inte bara kan omfatta sociologins uppfattning om vad som är socialt utan också kan visa på en fungerande modell av hur ett socialt system fungerar. Den uppfattningen understöds med hjälp av Niklas Luhmanns teori om sociala system och med hjälp av vygotsky-orienterade... (More) - Popular Abstract in Swedish
Avhandlingen diskuterar Sociologins position i debatten kring artificiell intelligens, d.v.s sådan AI som kan skapa kunskap på egen hand. Det påpekas att sociologin behöver göra en åtskillnad mellan två sätt att skapa AI system: Symbolisk AI (eller Klassisk AI) och Distribuerad AI – DAI.
Den sociologiska litteraturen visar en i stort kritisk attityd till symbolisk AI vilket är rätt. Avhandlingen handlar istället om att DAI inte bara kan omfatta sociologins uppfattning om vad som är socialt utan också kan visa på en fungerande modell av hur ett socialt system fungerar. Den uppfattningen understöds med hjälp av Niklas Luhmanns teori om sociala system och med hjälp av vygotsky-orienterade utbildningspedagoger som menar att de processer som leder till ny kunskap handlar om expansion och inte om problemlösning. Problemlösning kan bara handla om att laborera med existerande kunskap. Det påpekas dessutom att människan alltid har använt olika typer av artefakter för att kunna skapa kunskap. Vilket visar att människorna är beroende av sin miljö och att kunskap inte bara är något som finns i deras hjärnor utan också i deras kroppar och omgivningar.
Dessutom handlar AI inte om att robotar skall kunna tänka på samma sätt som människor utan om holistiska sociala processer. Stark AI behöver därför inte vara kapabel till en komplex funktionalitet som jämförs med människans tänkande. Detta ger en god möjlighet att förklara sociologins begrepp emergens – alltså hur sociala processer skapar kunskap utan att människor behöver vara inblandade.
Möjligheterna av AI måste utvärderas utifrån människornas kapaciteter att hantera reflexiva processer. Luhmann lär oss att vi måste skilja mellan tre olika former av reflexivitet: självreferens, reflexivitet och reflexion. För att kunna skapa kunskap i expanderande processer måste det finnas förutsättningar som tillåter reflexion. Vi kan finna stöd hos Luhmann för att de processer som skapar ny kunskap är beroende av en kapacitet för reflexion mellan de sociala systemen och deras omvärld. Stark AI har större kapacitet att hantera dessa processer än människor, eftersom det bästa argumentet för stark DAI är att de kan nå ett större område än människors hjärnor kan. Denna kapacitet för reflexion är ett bättre sätt att bedöma stark AI än kapacitet för kausalitet eller kapacitet för handling. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/1389611
- author
- Kåhre, Peter LU
- supervisor
- opponent
-
- Professor Ziemke, Tom, Institutionen för kognitionsvetenskap, Högskolan i Skövde
- organization
- alternative title
- On the Shoulders of AI-technology : Sociology of Knowledge and Strong Artificial Intelligence
- publishing date
- 2009
- type
- Thesis
- publication status
- published
- subject
- keywords
- James Wertsch, Gregory Bateson, Yrjö Engeström, Katherine N. Hayles, Lucy Suchman, Hubert Dreyfus, John Searle, David Bloor, Lev Vygotsky, Niklas Luhmann, Chinese room, Turing test, socionics, darwinism, emergence, relativism, posthumanism, environmentalism, situationism, social communication, second order cybernetics, systems theory, sociology of knowledge, connectionism, Strong artificial intelligence, distributed artificial intelligence
- in
- Lund Dissertations in Sociology
- volume
- 87
- pages
- 225 pages
- publisher
- Department of Sociology, Lund University
- defense location
- Kulturen
- defense date
- 2009-05-29 13:15:00
- ISSN
- 1403-6061
- ISBN
- 91-7267-289-7
- language
- Swedish
- LU publication?
- yes
- id
- fa3ce6cb-16ed-41b4-8537-0feaae50a774 (old id 1389611)
- date added to LUP
- 2016-04-01 14:18:17
- date last changed
- 2021-12-12 11:02:20
@phdthesis{fa3ce6cb-16ed-41b4-8537-0feaae50a774,
abstract = {{This dissertation is concerned with Sociology’s stance in the debate on Strong Artificial Intelligence, i.e. such AI that is able to shape new knowledge without human interference. There is a need for sociologists to realize the difference between two approaches to constructing AI systems: Symbolic AI (or Classic AI) and Distributed AI – DAI.<br/><br>
Sociological literature shows a largely critical attitude towards Symbolic AI, an attitude that is justified. The main theme of this dissertation is that DAI is not only compatible with Sociology’s approach to what is social, but also constitutes an apt model of how a social system functions. This is consolidated with help from Niklas Luhmann’s social systems theory and from Vygotsky-oriented education scientists who claim that processes leading to new knowledge are about expansion and not about problem solving. Problem solving only leads to elaborating existing knowledge. It is shown that human being has always used several types of artefacts and tools to produce culture and knowledge. This shows that humans are dependent on their environment and that knowledge is not only something that has to do with their brain, but also the rest of their bodies and their environments.<br/><br>
Further, Strong AI is not about robots thinking in the same way as humans, but about holistic social processes where the actor does not need to have a complex functionality. This provides a good opportunity to explain what sociologists call emergency, i.e. how social processes shape new knowledge independent of human actors.<br/><br>
<br/><br>
The possibility of AI has to be evaluated in terms of human beings’ capacities to cope with reflexive processes. Luhmann teaches us that we have to see the difference between three different forms of reflexivity: self-reference, reflexivity and reflection. We contend that, in order to be able to shape new knowledge in expanding processes, there must be circumstances that allow reflection. Luhmann writes that knowledge-producing processes are dependent on capacities for reflection between the social systems and their environments. Strong AI has more capacity to handle these processes than humans have, while the strongest argument for strong DAI is that it can reach a far wider area than human beings’ brains can. This capacity for reflection is a better way of judging the capacity of strong AI, than questions about causal capacity and capacity for action.}},
author = {{Kåhre, Peter}},
isbn = {{91-7267-289-7}},
issn = {{1403-6061}},
keywords = {{James Wertsch; Gregory Bateson; Yrjö Engeström; Katherine N. Hayles; Lucy Suchman; Hubert Dreyfus; John Searle; David Bloor; Lev Vygotsky; Niklas Luhmann; Chinese room; Turing test; socionics; darwinism; emergence; relativism; posthumanism; environmentalism; situationism; social communication; second order cybernetics; systems theory; sociology of knowledge; connectionism; Strong artificial intelligence; distributed artificial intelligence}},
language = {{swe}},
publisher = {{Department of Sociology, Lund University}},
school = {{Lund University}},
series = {{Lund Dissertations in Sociology}},
title = {{På AI-teknikens axlar : Om kunskapssociologin och stark artificiell intelligens}},
url = {{https://lup.lub.lu.se/search/files/3898263/1415449.pdf}},
volume = {{87}},
year = {{2009}},
}