ChatGPT passes Finnish Turing Test despite language challenges
Cultural Competence in Context: A Large Language Model Passes the Turing Test in Finland
Artificial Intelligence
Summary
The authors tested a version of ChatGPT in Finnish to see if people could tell it apart from humans in conversation. Even though Finnish is less represented in training data, the AI managed to pass the test. People often relied on casual language cues to guess if the speaker was human. The study suggests the Turing Test can show how well an AI fits into specific cultural and social contexts, not just its intelligence.
What this means in practice
- •For chatbot developers: Evaluate chatbot language and cultural adaptation in underrepresented languages using model-generated role prompting to improve realism and user trust.
- •For language technology teams: Use comparative Turing Tests to benchmark and improve AI models’ abilities to engage authentically in specific linguistic and cultural environments.
Authors
Otto Segersven, Pentti Henttonen
Abstract
We report the results of a Turing Test conducted in Finland in the Finnish language. Because languages and cultural contexts are unevenly represented in LLM training data, we expected the model (ChatGPT 5.2) to perform worse in a Finnish-language Turing Test than in previously studied English-language US contexts. We also present model-generated role prompting as a replicable technique for conducting comparative LLM-based Turing Tests designed to improve construct validity. Contrary to our expectations, the LLM passed the Finnish Turing Test. A prominent source of error was participants' reliance on linguistic cues, particularly colloquial Finnish, as markers of human authorship. We reframe the Turing Test from a test of intelligence to a comparative method for examining whether an AI system can display credible membership in a particular social world. Because its outcome reflects model capabilities, prompted identity, insider competence among human participants, and their AI literacy, the method provides a useful probe of the human-machine boundary across domains.