AnthroScore is an automatic metric designed to quantify the implicit anthropomorphism in language. It uses a masked language model to assess how non-human entities are framed as human by the surrounding context. The metric has been shown to correlate with human judgments of anthropomorphism and dimensions described in social science literature. The authors analyze 15 years of research papers and downstream news articles using AnthroScore, finding that anthropomorphism has increased over time, with papers related to language models having the highest levels of anthropomorphism. They also find that news headlines contain higher levels of anthropomorphism compared to research paper abstracts. The metric is lexicon-free and can be applied to a wide range of text sources, making it a valuable tool for understanding and mitigating misleading anthropomorphism in various contexts.AnthroScore is an automatic metric designed to quantify the implicit anthropomorphism in language. It uses a masked language model to assess how non-human entities are framed as human by the surrounding context. The metric has been shown to correlate with human judgments of anthropomorphism and dimensions described in social science literature. The authors analyze 15 years of research papers and downstream news articles using AnthroScore, finding that anthropomorphism has increased over time, with papers related to language models having the highest levels of anthropomorphism. They also find that news headlines contain higher levels of anthropomorphism compared to research paper abstracts. The metric is lexicon-free and can be applied to a wide range of text sources, making it a valuable tool for understanding and mitigating misleading anthropomorphism in various contexts.