När mänsklighet möter teknologi: En studie om sambandet mellan värme, kompetens och tillförlitlighet till ChatGPT
(2023) PSYK11 20231Department of Psychology
- Abstract (Swedish)
- AI-modellen ChatGPT har blivit omåttligt populär på kort tid och används dagligen av miljontals människor. Genom att analysera stora mängder data bestående av offentlig information utför AI-modellen tjänster som CVskrivande, diktkomponerande och besvarande av frågor om allt från vetenskap, till kultur och historia. ChatGPT används i mycket stor utsträckning, trots att den genererade informationen saknar källhänvisningar och referenser. Är vi naiva som blint litar på informationen ChatGPT förser oss med, och varför litar vi på just ChatGPT? Det som urskiljer ChatGPT från andra språkmodeller är det unika kommunikationssättet, som liknar det mänskliga. Denna studie undersöker huruvida modellen upplevs besitta mänskliga egenskaper i form av... (More)
- AI-modellen ChatGPT har blivit omåttligt populär på kort tid och används dagligen av miljontals människor. Genom att analysera stora mängder data bestående av offentlig information utför AI-modellen tjänster som CVskrivande, diktkomponerande och besvarande av frågor om allt från vetenskap, till kultur och historia. ChatGPT används i mycket stor utsträckning, trots att den genererade informationen saknar källhänvisningar och referenser. Är vi naiva som blint litar på informationen ChatGPT förser oss med, och varför litar vi på just ChatGPT? Det som urskiljer ChatGPT från andra språkmodeller är det unika kommunikationssättet, som liknar det mänskliga. Denna studie undersöker huruvida modellen upplevs besitta mänskliga egenskaper i form av värme och kompetens. Med bakgrund i Stereotype Content modellen och Humaness-value-loyalty modellen mäts deltagarnas upplevelse av mänsklighet hos ChatGPT, genom en digital enkätundersökning (N=55). Syftet med studien är att bidra med fördjupad kunskap kring samvariationen mellan värme, kompetens och tillförlitlighet. Arbetets resultat visar att ChatGPT upplevs utstråla både värme och kompetens. Dessutom påvisas en signifikant positiv korrelation mellan dimensionerna och tillförlitlighet. Vidare visar resultatet att endast kompetens är en signifikant prediktor till tillförlitlighet. Denna studie medvetandegör vikten av att ha samma källkritiska förhållningssätt till ChatGPT som till källor utan en interaktiv funktion, trots att mänskliga signaler kan påverka medvetandet till att uppleva informationen som tillförlitlig. (Less)
- Abstract
- The AI model ChatGPT has become immensely popular in a short time and is used daily by millions of people. By analyzing large amounts of data consisting of public information, the AI model performs services such as CV writing, composing poems and answering questions about everything from science, to culture and history. ChatGPT is used to a very large extent, despite the fact that the generated information lacks source references. Are we naive to blindly trust the information ChatGPT provides us with, and why do we trust ChatGPT in particular? What distinguishes ChatGPT from other language models is the unique human-like way of communicating. This study examines whether the model is perceived to possess human qualities in the form of... (More)
- The AI model ChatGPT has become immensely popular in a short time and is used daily by millions of people. By analyzing large amounts of data consisting of public information, the AI model performs services such as CV writing, composing poems and answering questions about everything from science, to culture and history. ChatGPT is used to a very large extent, despite the fact that the generated information lacks source references. Are we naive to blindly trust the information ChatGPT provides us with, and why do we trust ChatGPT in particular? What distinguishes ChatGPT from other language models is the unique human-like way of communicating. This study examines whether the model is perceived to possess human qualities in the form of warmth and competence. With a background in the Stereotype Content model and the Humaneness-value-loyalty model, the experience of humanity at ChatGPT is measured through a digital survey (N=55). The aim of the study was to contribute with in depth knowledge about the covariation between warmth, competence and reliability. The results of the work showed that ChatGPT is perceived to radiate both warmth and competence. In addition, a significant positive correlation between the dimensions and reliability was demonstrated. Furthermore, the results showed that only competence was a significant predictor of reliability. This study raises awareness of the importance of having the same source critical approach to ChatGPT as to sources without an interactive function, despite the fact that human cues can influence the mind to perceive the information as reliable. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9124471
- author
- Holster, Ella LU and Keifer, Johanna LU
- supervisor
- organization
- course
- PSYK11 20231
- year
- 2023
- type
- M2 - Bachelor Degree
- subject
- keywords
- artificiell intelligens, ChatGPT, SCM, HVL, tillförlitlighet, värme, kompetens, artificial intelligence, reliability, warmth, competence
- language
- Swedish
- id
- 9124471
- date added to LUP
- 2023-06-16 08:27:28
- date last changed
- 2023-06-16 08:27:28
@misc{9124471, abstract = {{The AI model ChatGPT has become immensely popular in a short time and is used daily by millions of people. By analyzing large amounts of data consisting of public information, the AI model performs services such as CV writing, composing poems and answering questions about everything from science, to culture and history. ChatGPT is used to a very large extent, despite the fact that the generated information lacks source references. Are we naive to blindly trust the information ChatGPT provides us with, and why do we trust ChatGPT in particular? What distinguishes ChatGPT from other language models is the unique human-like way of communicating. This study examines whether the model is perceived to possess human qualities in the form of warmth and competence. With a background in the Stereotype Content model and the Humaneness-value-loyalty model, the experience of humanity at ChatGPT is measured through a digital survey (N=55). The aim of the study was to contribute with in depth knowledge about the covariation between warmth, competence and reliability. The results of the work showed that ChatGPT is perceived to radiate both warmth and competence. In addition, a significant positive correlation between the dimensions and reliability was demonstrated. Furthermore, the results showed that only competence was a significant predictor of reliability. This study raises awareness of the importance of having the same source critical approach to ChatGPT as to sources without an interactive function, despite the fact that human cues can influence the mind to perceive the information as reliable.}}, author = {{Holster, Ella and Keifer, Johanna}}, language = {{swe}}, note = {{Student Paper}}, title = {{När mänsklighet möter teknologi: En studie om sambandet mellan värme, kompetens och tillförlitlighet till ChatGPT}}, year = {{2023}}, }