{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/semantics-derived-automatically-from-language","title":"Semantics derived automatically from language corpora contain human-like biases","arxiv_id":"1608.07187","date":"2016-08-25","proceeding":null,"authors":["Aylin Caliskan","Joanna J. Bryson","Arvind Narayanan"],"abstract":"Artificial intelligence and machine learning are in a period of astounding\ngrowth. However, there are concerns that these technologies may be used, either\nwith or without intention, to perpetuate the prejudice and unfairness that\nunfortunately characterizes many human institutions. Here we show for the first\ntime that human-like semantic biases result from the application of standard\nmachine learning to ordinary language---the same sort of language humans are\nexposed to every day. We replicate a spectrum of standard human biases as\nexposed by the Implicit Association Test and other well-known psychological\nstudies. We replicate these using a widely used, purely statistical\nmachine-learning model---namely, the GloVe word embedding---trained on a corpus\nof text from the Web. Our results indicate that language itself contains\nrecoverable and accurate imprints of our historic biases, whether these are\nmorally neutral as towards insects or flowers, problematic as towards race or\ngender, or even simply veridical, reflecting the {\\em status quo} for the\ndistribution of gender with respect to careers or first names. These\nregularities are captured by machine learning along with the rest of semantics.\nIn addition to our empirical findings concerning language, we also contribute\nnew methods for evaluating bias in text, the Word Embedding Association Test\n(WEAT) and the Word Embedding Factual Association Test (WEFAT). Our results\nhave implications not only for AI and machine learning, but also for the fields\nof psychology, sociology, and human ethics, since they raise the possibility\nthat mere exposure to everyday language can account for the biases we replicate\nhere.","url_abs":"http://arxiv.org/abs/1608.07187v4","url_pdf":"http://arxiv.org/pdf/1608.07187v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"semantics-derived-automatically-from-language","repo_url":"https://github.com/YunjinPark/modu_project","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"machine-learning","task_name":"BIG-bench Machine Learning"},{"task_slug":"ethics","task_name":"Ethics"},{"task_slug":"sociology","task_name":"Sociology"}],"methods":[{"method_slug":"glove","method_name":"GloVe"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1608.07187","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}