Anthropomorphism, the attribution of human-like characteristics to non-human entities, is a common phenomenon in technology discussions. ANTHROSCORE is an automatic metric that quantifies implicit anthropomorphism in language using a masked language model to assess how non-human entities are framed as human in the surrounding context. It aligns with human judgments and social science definitions of anthropomorphism. The study analyzes 15 years of research papers and downstream news articles, revealing that anthropomorphism has steadily increased over time, with language model-related papers showing the highest levels. Within ACL papers, increases in anthropomorphism correlate with key neural advancements. News headlines show higher anthropomorphism than research paper abstracts. ANTHROSCORE is lexicon-free, making it applicable to various text sources. The study highlights the risks of misleading anthropomorphism in technology discourse, including overestimating capabilities, over-reliance on technology, and corporate responsibility avoidance. However, anthropomorphism also has benefits, such as facilitating learning and fostering environmentalism. The study provides recommendations for authors and the broader community to minimize misleading anthropomorphism. ANTHROSCORE can be applied to various contexts beyond research papers and news articles, including corporate branding, conspiracy theories, and relationships with pets and objects. The study acknowledges limitations, including its focus on English data and potential biases in pre-trained language models. Overall, ANTHROSCORE offers a tool for analyzing anthropomorphism across different cultures, languages, and times.Anthropomorphism, the attribution of human-like characteristics to non-human entities, is a common phenomenon in technology discussions. ANTHROSCORE is an automatic metric that quantifies implicit anthropomorphism in language using a masked language model to assess how non-human entities are framed as human in the surrounding context. It aligns with human judgments and social science definitions of anthropomorphism. The study analyzes 15 years of research papers and downstream news articles, revealing that anthropomorphism has steadily increased over time, with language model-related papers showing the highest levels. Within ACL papers, increases in anthropomorphism correlate with key neural advancements. News headlines show higher anthropomorphism than research paper abstracts. ANTHROSCORE is lexicon-free, making it applicable to various text sources. The study highlights the risks of misleading anthropomorphism in technology discourse, including overestimating capabilities, over-reliance on technology, and corporate responsibility avoidance. However, anthropomorphism also has benefits, such as facilitating learning and fostering environmentalism. The study provides recommendations for authors and the broader community to minimize misleading anthropomorphism. ANTHROSCORE can be applied to various contexts beyond research papers and news articles, including corporate branding, conspiracy theories, and relationships with pets and objects. The study acknowledges limitations, including its focus on English data and potential biases in pre-trained language models. Overall, ANTHROSCORE offers a tool for analyzing anthropomorphism across different cultures, languages, and times.