{"id":125690,"date":"2021-05-19T18:17:00","date_gmt":"2021-05-19T18:17:00","guid":{"rendered":"https:\/\/precoinnews.com\/?p=125690"},"modified":"2021-05-19T18:17:00","modified_gmt":"2021-05-19T18:17:00","slug":"twitter-finds-its-ai-tends-to-crop-out-black-people-men-from-photos","status":"publish","type":"post","link":"https:\/\/precoinnews.com\/markets\/twitter-finds-its-ai-tends-to-crop-out-black-people-men-from-photos\/","title":{"rendered":"Twitter finds its AI tends to crop out Black people, men from photos"},"content":{"rendered":"
(Reuters) – Twitter Inc\u2019s image-cropping algorithm has a problematic bias toward excluding Black people and men, the company said in new research on Wednesday, adding that \u201chow to crop an image is a decision best made by people.\u201d<\/p> The study by three of its machine learning researchers was conducted after user criticism last year about image previews in posts excluding Black people\u2019s faces.<\/p>\n It found an 8% difference from demographic parity in favor of women, and a 4% favor toward white individuals.<\/p>\n The paper cited several possible reasons, including issues with image backgrounds and eye color, but said none were an excuse.<\/p>\n \u201cMachine learning based cropping is fundamentally flawed because it removes user agency and restricts user\u2019s expression of their own identity and values, instead imposing a normative gaze about which part of the image is considered the most interesting,\u201d the researchers wrote.<\/p>\n To counter the problem, Twitter recently started showing standard aspect ratio photos in full – without any crop – on its mobile apps and is trying to expand that effort.<\/p>\n The researchers also assessed whether crops favored women\u2019s bodies over heads, reflecting what is known as the \u201cmale gaze,\u201d but found that does not appear to be the case.<\/p>\n The findings are another example of the disparate impact from artificial intelligence systems including demographic biases identified in facial recognition and text analysis, the paper said.<\/p>\n Work by researchers at Microsoft Corp and the Massachusetts Institute of Technology in 2018 and a later U.S. government study found that facial analysis systems misidentify people of color more often than white people.<\/p>\n Amazon Inc in 2018 scrapped an AI recruiting tool that showed bias against women.<\/p>\n