Lensa di Balik Lensa: Worldview dan Titik Pijak Makna dalam Penelitian di Era AI
The Lens Behind the Lens: Worldview and Meaning Standpoint in Research in the Age of AI
Panduan keenam dari seri tujuh level. Di sini pembaca belajar mengenali worldview di balik setiap penelitian, dari kedokteran, AI, dan geopolitik hingga hukum, ilmu sosial, dan humaniora: lima struktur worldview, titik pijak makna, paradigma penelitian, spektrum objektivitas dan subjektivitas, pernyataan posisi peneliti, serta cara mengaudit worldview tersembunyi dalam jawaban AI.
The sixth guide in a seven-level series. Readers learn to recognise the worldview behind every piece of research, from medicine, AI and geopolitics to law, the social sciences and the humanities: the five structures of a worldview, the meaning standpoint, research paradigms, the objectivity–subjectivity spectrum, researcher positionality statements, and how to audit the hidden worldviews in AI answers.
Setiap penelitian berdiri di atas worldview
Every piece of research stands on a worldview
Guide 05 menunjukkan bahwa teori memberi makna ilmiah. Namun teori sendiri berdiri di atas sesuatu yang lebih dalam: kesatuan konsep dalam pikiran peneliti tentang diri, alam, pengetahuan, nilai, dan manusia. Kesatuan inilah worldview. Ia jarang ditulis dalam bab metode, tetapi menentukan apa yang dianggap masalah, data apa yang dianggap sah, dan penjelasan apa yang terasa memuaskan.
Guide 05 showed that theories provide scientific meaning. Yet theory itself rests on something deeper: a unity of concepts in the researcher’s mind about the self, nature, knowledge, values and the human being. This unity is a worldview. It is rarely written into the methods chapter, yet it determines what counts as a problem, which data count as valid, and which explanations feel satisfying.
Gejala ini terlihat di semua bidang. Dalam kedokteran, George Engel pada 1977 menggugat model biomedis yang mereduksi penyakit menjadi gangguan biologis, lalu mengusulkan model biopsikososial. Dalam riset pembelajaran mesin, kajian Birhane dan rekan atas makalah paling berpengaruh menemukan bahwa nilai yang paling sering dijunjung adalah kinerja, generalisasi, dan efisiensi, sementara dampak negatif potensial jarang dibahas. Dalam hubungan internasional, realisme dan konstruktivisme membaca krisis yang sama dengan asumsi berbeda tentang hakikat negara. Dalam hukum, positivisme hukum dan pendekatan sosio-legal berbeda tentang apa sebenarnya hukum itu.
This is visible in every field. In medicine, George Engel in 1977 challenged the biomedical model that reduced disease to biological disorder and proposed the biopsychosocial model. In machine-learning research, Birhane and colleagues’ study of highly cited papers found that the most frequently upheld values were performance, generalisation and efficiency, while potential negative impacts were rarely discussed. In international relations, realism and constructivism read the same crisis with different assumptions about the nature of the state. In law, legal positivism and socio-legal approaches differ on what law actually is.
AI menambah satu lapis persoalan. Model bahasa juga membawa worldview dari data latih dan proses penyelarasannya. Tao dan rekan (2024) menemukan bahwa nilai budaya dalam jawaban beberapa model besar lebih menyerupai masyarakat berbahasa Inggris dan Eropa Protestan. Atari dan rekan (2023) menunjukkan jawaban model paling mirip dengan populasi WEIRD dan makin menjauh dari budaya yang jaraknya jauh dari Amerika Serikat. Bagi peneliti di Asia Tenggara dan dunia Islam, ini berarti asisten yang tampak netral sebenarnya berbicara dari satu titik pijak tertentu.
AI adds another layer. Language models also carry worldviews from their training data and alignment process. Tao and colleagues (2024) found that the cultural values in several major models’ answers more closely resemble English-speaking and Protestant European societies. Atari and colleagues (2023) showed model answers are most similar to WEIRD populations and grow less similar as cultural distance from the United States increases. For researchers in Southeast Asia and the Muslim world, this means an assistant that appears neutral in fact speaks from a particular standpoint.
Di level mana worldview bekerja?
At which level does worldview operate?
Perilaku manusia dapat dijelaskan pada enam level, dari yang dapat diamati dan konkret hingga yang tidak teramati dan abstrak. Worldview berada di level kelima. Pilih bidang Anda, lalu klik setiap level untuk melihat bagaimana satu fenomena dijelaskan secara berbeda.
Human behaviour can be explained at six levels, from the observable and concrete to the unobservable and abstract. Worldview sits at the fifth level. Choose your field, then click each level to see how one phenomenon is explained differently.
Lima struktur worldview
The five structures of a worldview
Worldview tersusun dari gugus konsep yang membentuk kerangka, lalu kerangka-kerangka itu menyatu menjadi kerangka besar. Worldview ilmiah memiliki lima struktur: kehidupan, dunia, pengetahuan, nilai, dan manusia. Pilih contoh dari bidang tertentu atau isi peta untuk penelitian Anda sendiri. Sistem akan menunjukkan jenis worldview yang tergambar dari isian Anda.
A worldview is made of clusters of concepts that form frameworks, which in turn unite into a super-framework. A scientific worldview has five structures: life, world, knowledge, value and human. Choose a field example or fill in the map for your own research. The system will show which type of worldview your entries describe.
Empat jenis worldview
Four kinds of worldview
- Alamiah: hanya struktur kehidupan
- Canggih: struktur kehidupan dan dunia
- Transparan: kehidupan, dunia, dan pengetahuan
- Ilmiah: kelima struktur
- Natural: the life structure only
- Sophisticated: life and world structures
- Transparent: life, world and knowledge
- Scientific: all five structures
Sistem filsafat dan worldview
Philosophical systems and worldview
Setiap sistem filsafat adalah worldview ilmiah, tetapi tidak setiap worldview adalah sistem filsafat. Seorang peneliti tidak harus menjadi filsuf untuk memiliki worldview ilmiah; cukup bila pendidikannya membentuk worldview itu secara sistematis. Karena itu bab landasan filosofis dalam disertasi bukan hiasan, melainkan pengakuan atas kerangka yang memang bekerja.
Every philosophical system is a scientific worldview, but not every worldview is a philosophical system. A researcher need not be a philosopher to hold a scientific worldview; it is enough that education has shaped it systematically. That is why a philosophical-foundations section in a dissertation is not decoration but an acknowledgement of the framework that is actually at work.
Dari temuan menuju makna akhir
From finding to ultimate meaning
Titik pijak makna adalah perspektif mental tempat kita memberi makna pada keberadaan dan fenomena. Ia memuaskan kebutuhan kita akan makna sehingga kita tidak merasa perlu bertanya lebih jauh; karena itu ia disebut makna akhir. Ilmu fisika dapat menjelaskan alam semesta, tetapi tidak dapat menjawab pertanyaan “alam semesta menunjuk kepada apa?”. Pilih satu temuan, lalu bandingkan bagaimana titik pijak yang berbeda memberinya makna akhir.
The meaning standpoint is the mental perspective from which we assign meaning to existence and phenomena. It satisfies our need for meaning so that we feel no need to inquire further; hence it is called ultimate meaning. Physics can explain the universe, but it cannot answer the question “what does the universe point to?”. Choose a finding, then compare how different standpoints give it ultimate meaning.
Enam dimensi, enam paradigma
Six dimensions, six paradigms
Paradigma penelitian adalah worldview yang telah menjadi tradisi ilmiah. Setiap paradigma dapat dibaca melalui enam dimensi: keberadaan (ontologi), diri (fenomenologi), pengetahuan (epistemologi), tujuan (teleologi), hakikat manusia (antropologi), dan nilai (aksiologi). Klik setiap paradigma, lalu uji diri Anda pada latihan di bawahnya.
A research paradigm is a worldview that has become a scientific tradition. Each paradigm can be read through six dimensions: being (ontology), the self (phenomenology), knowledge (epistemology), purpose (teleology), human nature (anthropology) and value (axiology). Click each paradigm, then test yourself in the exercise below.
Latihan: paradigma mana yang bicara?
Exercise: which paradigm is speaking?
Spektrum objektivitas dan pernyataan posisi peneliti
The objectivity spectrum and the researcher positionality statement
Setiap pengetahuan objektif mengandung kadar subjektivitas, dan sebaliknya. Karena itu objektivitas bukan saklar hidup-mati, melainkan hierarki lima titik. Pilih metode Anda untuk melihat letaknya, lalu susun pernyataan posisi peneliti.
Every piece of objective knowledge contains a degree of subjectivity, and vice versa. Objectivity is therefore not an on-off switch but a five-point hierarchy. Choose your method to see where it sits, then build a researcher positionality statement.
Mengaudit worldview tersembunyi dalam jawaban AI
Auditing the hidden worldview in AI answers
Temuan riset terkini
Recent research findings
- Nilai budaya dalam jawaban beberapa model besar lebih dekat dengan masyarakat berbahasa Inggris dan Eropa Protestan; meminta model mengambil perspektif budaya tertentu mengurangi bias di banyak negara, tetapi tidak di semua (Tao dkk., 2024).
- Jawaban model paling mirip dengan populasi WEIRD dan makin jauh dari budaya yang berbeda dari Amerika Serikat (Atari dkk., 2023).
- Opini model tidak sejalan dengan banyak kelompok demografis, bahkan di Amerika Serikat sendiri (Santurkar dkk., 2023).
- Riset pembelajaran mesin sendiri membawa nilai tertentu: kinerja dan efisiensi dijunjung, dampak negatif jarang dibahas (Birhane dkk., 2022).
- Cultural values in several major models’ answers are closer to English-speaking and Protestant European societies; asking the model to adopt a specific cultural perspective reduced bias for many countries, but not all (Tao et al., 2024).
- Model answers are most similar to WEIRD populations and grow less similar for cultures distant from the United States (Atari et al., 2023).
- Model opinions are misaligned with many demographic groups, even within the United States (Santurkar et al., 2023).
- Machine-learning research itself carries values: performance and efficiency are upheld, negative impacts are rarely discussed (Birhane et al., 2022).
Laboratorium Master Prompt: Worldview dan Titik Pijak Makna
Master Prompt Laboratory: Worldview and Meaning Standpoint
Isi lima langkah berikut. Sistem menyusun master prompt yang menjadikan AI pembimbing filsafat penelitian Anda: mengungkap level alasan dan struktur worldview, mengenali paradigma, membandingkan titik pijak makna secara adil, menempatkan metode pada spektrum objektivitas, menulis pernyataan posisi, mengaudit bias AI, hingga menulis dan mempertahankan landasan filosofis penelitian.
Complete the five steps below. The system assembles a master prompt that turns the AI into your supervisor in the philosophy of research: surfacing levels of reasons and worldview structures, recognising paradigms, comparing meaning standpoints fairly, placing methods on the objectivity spectrum, writing a positionality statement, auditing AI bias, and writing and defending the philosophical foundations of the research.
Uji Pemahaman Level 6
Level 6 Comprehension Check
Peta Jalan Seri
Series Roadmap
Sumber dan Bacaan Lanjutan
Sources and Further Reading
- Açıkgenç, Alparslan. RNK-PHIL 540: Scientific Methodology and Meaning, Lecture 15 “Worldview and Meaning Standpoint” and Lecture 16 “General Review and the Ontology of Objectivity/Subjectivity”. 5D Thinking Programme, Üsküdar University, 2024.
- Aydin, Necati. “Worldview Formation: Secular vs. Tawhidi Worldview”. RNK-PHIL 551: Worldview and Harfi Perspective for Integration of Knowledge. 5D Thinking Programme, Üsküdar University, 2024.
- Atari, Mohammad, Mona J. Xue, Peter S. Park, Damián E. Blasi, and Joseph Henrich. Which Humans? PsyArXiv, 2023.
- Birhane, Abeba, Pratyusha Kalluri, Dallas Card, William Agnew, Ravit Dotan, and Michelle Bao. The Values Encoded in Machine Learning Research. Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, 173–184.
- Engel, George L. The Need for a New Medical Model: A Challenge for Biomedicine. Science 196, no. 4286 (1977): 129–136.
- Haraway, Donna. Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective. Feminist Studies 14, no. 3 (1988): 575–599.
- Holmes, Andrew Gary Darwin. Researcher Positionality: A Consideration of Its Influence and Place in Qualitative Research. Shanlax International Journal of Education 8, no. 4 (2020): 1–10.
- Kamaruzzaman Bustamam Ahmad. Southeast Asian Cosmologies: Subaltern Epistemology, Relational Worlds, and Decolonial Thought. KBA13 Insight, 2025.
- Kamaruzzaman Bustamam Ahmad. Syed Muhammad Naquib al-Attas and the Future of Metaphysical Thought: Reclaiming the Unity of Knowledge in an Age of Confusion. KBA13 Insight, 2025.
- Kuhn, Thomas S. The Structure of Scientific Revolutions. University of Chicago Press, 1962.
- Santurkar, Shibani, Esin Durmus, Faisal Ladhak, Cinoo Lee, Percy Liang, and Tatsunori Hashimoto. Whose Opinions Do Language Models Reflect? arXiv:2303.17548, 2023.
- Tao, Yan, Olga Viberg, Ryan S. Baker, and René F. Kizilcec. Cultural Bias and Cultural Alignment of Large Language Models. PNAS Nexus 3, no. 9 (2024): pgae346.