University of Arts in Belgrade
Digital Repository of the Institute for Philosophy and Social Theory, University of BelgradeNot a member yet
4571 research outputs found
Sort by
Does GPT-4 surpass human performance in linguistic pragmatics?
No full textAs Large Language Models (LLMs) become increasingly integrated into everyday life as general-purpose multimodal AI systems, their capabilities to simulate human understanding are under examination. This study investigates LLMs’ ability to interpret linguistic pragmatics, which involves context and implied meanings. Using Grice’s communication principles, we evaluated both LLMs (GPT-2, GPT-3, GPT-3.5, GPT-4, and Bard) and human subjects (N = 147) on dialogue-based tasks. Human participants included 71 primarily Serbian students and 76 native English speakers from the United States. Findings revealed that LLMs, particularly GPT-4, outperformed humans. GPT-4 achieved the highest score of 4.80, surpassing the best human score of 4.55. Other LLMs performed well: GPT-3.5 scored 4.10, Bard 3.75, and GPT-3 3.25; GPT-2 had the lowest score of 1.05. The average LLM score was 3.39, exceeding the human cohorts’ averages of 2.80 (Serbian students) and 2.34 (U.S. participants). In the ranking of all 155 subjects (including LLMs and humans), GPT-4 secured the top position, while the best human ranked second. These results highlight significant progress in LLMs’ ability to simulate understanding of linguistic pragmatics. Future studies should confirm these findings with more dialogue-based tasks and diverse participants. This research has important implications for advancing general-purpose AI models in various communication-centered tasks, including potential application in humanoid robots in the future
Human Doctors Versus AI Models: Advice Perceptions in Serbia, Kazakhstan, and Kyrgyzstan
No full textThis chapter examines perceptions among individuals from Serbia, Kazakhstan, and Kyrgyzstan regarding the trustworthiness, usefulness, and empathy of medical advice provided by AI models compared to that offered by human doctors. A mixed-methods approach was employed, involving online surveys where 442 participants rated medical advice from three AI models (GPT-4o, Llama 3.1, Gemini Pro 1.5) and a real doctor across three medical scenarios. Participants evaluated each piece of advice on a 7-point Likert scale for trustworthiness, usefulness, and empathy. They indicated whether they believed the advice was generated by AI or provided by a real doctor. Results showed that participants across all three countries rated advice from human doctors significantly higher in trustworthiness, usefulness, and empathy than AI-generated advice. Familiarity with AI technology and trust in AI-generated information was highest in Kazakhstan, correlating with greater acceptance and usage of AI for health queries. Participants exhibited moderate ability to distinguish between AI and human advice, suggesting that AI models are advancing in producing human-like communication. The study concludes that although AI models show promise in generating medical advice, the human elements of trust and empathy remain indispensable in healthcare
Kritički dijalog s Metzingerovom filozofijom čiste svijesti
No full textČista svijest, kao neegoistična svijest bez posebnih intencionalnih sadržaja i fenomenalnih
kvaliteta, istaknuta je tema znanstvenih istraživanja i filozofije od šezdesetih godina prošlog
stoljeća, ali se u posljednjih trideset godina pažnja ovoj temi višestruko povećala. Kulminacija je ovog procesa knjiga Thomasa Metzingera The Elephant and the Blind, objavljena
2024. godine, koja se oslanja na oko 1400 izvještaja u svrhu predstavljanja nove filozofije
svijesti, uzimajući u obzir kako duhovne tradicije, tako i suvremene znanstvene uvide. Nakon kratkog izlaganja glavnih teza Metzingerove knjige, ovaj članak otvara nekoliko pitanja
u okviru kritičkog dijaloga. S jedne strane, treba li meditaciju nazvati »kraljevskim putem«
prema čistoj svijesti, a s druge strane, treba li nasuprot tome više pažnje posvetiti afektivnim
putevima prema čistoj svijesti, prije svega određenim tradicijama i praksama misticizma.
Istovremeno, imajući u vidu tezu da je svijest inherentno samosvjesna, članak sugerira da
bi čistu samosvijest trebalo razumjeti kao samoafektivnos
Dark minds do not like animals: the dark tetrad traits as predictors of relations with non-human animals
No full textLithium Mining and Environmental (In)justice
No full textThis publication was produced as part of the project ‘‘Environmental
(In)justice: Project Jadar and Lithium Frenzy,’’ within the international
initiative ‘‘Global Justice and Humanities Practices’’ of the Consortium
of Humanities Centers and Institutes (CHCI). The aim of the initiative
was to explore how the humanities and social sciences can respond to
global challenges by connecting theoretical reflection with concrete
social problems.1 The project was conducted by the research team
of the Institute for Philosophy and Social Theory at the University of
Belgrade, in collaboration with interested researchers from Serbia
and abroad. The research focus was directed toward the connection
between lithium mining, environmental degradation, community displacement,
and undemocratic patterns of governance, with particular
emphasis on the controversial Jadar project in Serbia
Ko smo mi i šta možemo zajedno: identifikacija sa studentima u blokadi i ocena grupne efikasnosti predviđaju učešće u studentskim protestima/blokadama
No full textСтудентске блокаде/протести представљају пример колективне акције, која је усмерена на
добробит групе и шире заједнице. Иако су ове акције од широког значаја, ипак не учествују сви
студенти у њима у једнакој мери. Према доминантном моделу у литератури, социјалноидентитетском
моделу колективне акције (Social Identity Model of Collective Action, SIMCA), један од кључних
предуслова ангажмана је да се „активира“ групни идентитет који је доживљен као релевантан и важан.
Тек када доживљава овакву идентификацију, особа ће опажати да припада групи у чију корист се акција
спроводи, то јест доживљавати да „ми“ спроводимо неку акцију. Друга кључна детерминанта акције је
процена групне ефикасности – уколико особа очекује да група има капацитет да изведе акције и да оне
резултирају неким променама, утолико ће бити спремнија да се у њима и ангажује. На узорку од 504
студената (69.6% жена) свих универзитета у Србији тестирали смо ове претпоставке. Меру
идентификације са студентима у блокади формирали смо на основу процене осећаја заједништва са
осталим студентима у блокади као и доживљаја да учешће у блокади постаје важно за њих као особу
(α = .75). Учесници су такође оценили у ком степену очекују да студенти могу да буду успешни у
акцијама и да доведу до промена у друштву, на основу чега смо извели меру групне ефикасности (α =
.90). Проверили смо како ове варијабле предвиђају три исхода: учешће у студентским протестним
акцијама, учешће у пленуму и намеру да се настави учешће у блокадама. Рeгресиона анализа показује
да предиктори објашњавају 6.9% варијансе учешћа у протестним акцијама (F (2, 501) = 17.78, p < .001),
при чему је идентификација једини значајан предиктор (β = .25, t (501) = 14.21, p < .001). Слично, модел
објашњава 4.5% варијансе учешћа у пленумима (F (2, 501) = 12.73, p < .001), са идентификацијом као
значајним предиктором (β = .22, t (501) = 8.9, p < .001). Најзад, идентификација са студентима у блокади
(β = .15, t (501) = 2.71, p = .007), као и групна ефикасност (β = .32, t (501) = 5.81, p < .001) дају највећи
допринос предвиђању спремности на даље учешће у акцијама (19.7% варијансе, F (2, 501) = 61.60, p <
.001). Резултати подржавају претпоставке социјалноидентитетског модела колективне акције,
потврђујући да је снага привржености групи студената у блокади значајна детерминанта актуелног
учешћа у студентским протестима/блокадама, док је оцена групне ефикасности предиктивнија за
будуће учешће
The One and Many. Reception of the Neoplatonic Notion of Hierarchy in 14th and 15th Century Balkans
No full textThe Gender Gap in the Care Economy Is Larger in Highly Developed Countries: Sociocultural Explanations for Paradoxical Findings
No full textDespite the growing demand for care economy employees (e.g., nurses, teachers, and social workers), men remain underrepresented in these careers. While economically developed countries support more equal rights for women and men, their labor markets are highly gender segregated (Charles, 1992, 2003). We conducted a focused investigation of this paradoxical pattern in the care economy, testing whether gender gaps in care economy career interest are larger in more economically developed countries and, if so, what psychological and cultural factors underlie these patterns. We examined these questions with labor data from 70 countries (Study 1) and a preregistered study of career interests among 19,240 university students from 49 countries (Study 2). Although more economically developed countries tend to promote greater gender equality, our results reveal that the gender gap in care economy representation (Study 1) and interest (Study 2) is especially large in such countries. We did not observe parallel patterns for science, technology, engineering, and mathematics representation or interest. Results from Study 2 supported an integrated theoretical account of this development paradox in care economy interest: Cross-national variation in the gender gap in care economy interest was predicted by country-level variation in economic development and individualism/collectivism but not by self-expression values or country-level gender equality, countering prior (controversial) claims of a gender equality paradox. Furthermore, larger gender gaps in communal values (e.g., men’s lower valuing of helping and caring for others) were a proximal predictor of larger gender gaps in care economy interest in highly economically developed countries.broj koautora 13
