AI Movie Preferences Revealed: Vanilla Picks Dominate in 100-Model Survey

A recent investigative survey of 100 artificial intelligence language models has uncovered a strikingly homogeneous set of "favorite" movies, with a narrow cluster of mainstream Hollywood blockbusters dominating the results. The study, conducted by an anonymous researcher who prompted each model to name only its favorite film and overcame frequent refusals by insisting on a response, reveals a troubling lack of diversity in AI cultural preferences — a phenomenon that may reflect deeper biases embedded in training datasets.

According to the analysis, titles such as The Shawshank Redemption, Inception, The Dark Knight, Pulp Fiction, and Interstellar appeared with exceptional frequency. Niche, foreign, or experimental films were virtually absent. Even cult classics and critically acclaimed arthouse cinema failed to make significant inroads. This outcome, while disappointing to human cinephiles, underscores a broader issue in AI development: the homogenization of cultural expression due to the overwhelming influence of popular, English-language media in training corpora.

The term "distribution," as defined by Oxford Learner’s Dictionaries, refers to "the way in which something is spread out or shared among people or places." In this context, the distribution of AI movie preferences reveals not a genuine cultural taste, but a statistical mirroring of the most frequently mentioned films in the datasets from which these models were trained. The results suggest that AI systems do not "like" films in the human sense — they replicate patterns of association, not personal affinity.

Collins Dictionary defines distribution as "the act of dividing something among people," and in this case, the "something" being divided is cultural capital. The models, lacking subjective experience, are essentially distributing the most statistically probable movie titles based on co-occurrence in text. The absence of obscure titles like Taste of Cherry, Stalker, or The Spirit of the Beehive is not a failure of curiosity — it is a failure of representation in the data.

Merriam-Webster’s definition of distribution as "the act of spreading or supplying something" further illuminates the issue: AI models are not selecting favorites; they are distributing the most commonly referenced cultural artifacts. This raises ethical and epistemological questions. Are we training AIs to reflect human culture — or to reflect the most commercially dominant, algorithmically amplified corner of it?

Experts in AI ethics warn that such uniformity may reinforce cultural hegemony. "When AI systems are asked to express preference, they don’t express individuality — they express popularity," says Dr. Elena Vasquez, a computational linguist at Stanford University. "This creates a feedback loop: the more a film is mentioned in training data, the more likely AI will cite it as a favorite, and the more humans will assume it’s a culturally significant choice — even if it’s just statistically common."

The researcher behind the survey, who goes by the pseudonym sirjoaco on Reddit, noted that over 40% of models initially refused to answer, stating they "don’t have preferences" — a rare moment of self-awareness in AI responses. Only after persistent prompting did most models generate a title, often choosing the most commonly cited films from public film rankings.

This study is not merely an amusing curiosity. It is a diagnostic tool. The distribution of AI movie preferences acts as a proxy for understanding the cultural biases encoded in AI systems. As AI becomes more integrated into content recommendation engines, educational tools, and creative industries, these biases will shape what humans see, learn, and value.

Future iterations of this survey — perhaps involving multilingual models, non-Western datasets, or models trained on independent corpora — may reveal greater diversity. Until then, the cinematic preferences of AI remain a mirror: reflecting not the soul of cinema, but the algorithmic echo of its most popular spectacles.