Reference:

Timo Honkela. Semioottisesti kyvykkäät koneet sovelluksineen. Futura, 31(2):27–41, 2012.

Abstract:

In this article, the future of artificial intelligence is considered from an interdisciplinary point of view. A scenario in which machines that can learn to understand in a contextually appropriate manner is presented. These kinds of machines are called to be semiotically competent. This includes an ability to learn and refine multimodally grounded and contextually relevant conceptual systems and to conduct meaning negotiations. Statistical machine learning, pattern recognition and probabilistic modeling form a methodological basis for such developments. Grounded Intersubjective Concept Analysis (GICA) is presented as an example of methodology that paves way to the presented scenario. The GICA method can be used to analyze similarities and differences in understanding regarding different individuals and contexts. Regarding these kinds of developments in artificial intelligence, different potential applications in and implications for education, science and society in general as well as developments in the health and wellbeing area and environmental domain are discussed. The main conclusion is that there are many possibilities for positive outcomes and deliberating consequences, opposite to the dark views often presented in association with intelligent technologies.

Suggested BibTeX entry:

@article{HonkelaFutura12,
    author = {Timo Honkela},
    journal = {Futura},
    language = {fin},
    number = {2},
    pages = {27-41},
    title = {Semioottisesti kyvykk{\"a}{\"a}t koneet sovelluksineen},
    volume = {31},
    year = {2012},
}