{"resolved":true,"match_type":"uuid","source_type":"original_book","book_id":"56fa816e-042b-40dd-9a08-6bc236c079d8","title":"人工智能政治哲学","author":null,"edition":null,"publisher":null,"year":null,"isbn":"9787208189485","requested_by":"printed_page","book":{"source_type":"original_book","book_id":"56fa816e-042b-40dd-9a08-6bc236c079d8","id":"56fa816e-042b-40dd-9a08-6bc236c079d8","uuid":"56fa816e-042b-40dd-9a08-6bc236c079d8","title":"人工智能政治哲学","normalized_title":"人工智能政治哲学","author":null,"edition":null,"publisher":null,"year":null,"isbn":"9787208189485","isbns":["9787208189485"],"parsed":true,"ai_readable":true,"document_type":"book","page_count":261,"navigation_url":"https://workshop.closeclaw.top/ai/book/56fa816e-042b-40dd-9a08-6bc236c079d8","title_url":"https://workshop.closeclaw.top/ai/title/%E4%BA%BA%E5%B7%A5%E6%99%BA%E8%83%BD%E6%94%BF%E6%B2%BB%E5%93%B2%E5%AD%A6","fixed_page_url_template":"https://workshop.closeclaw.top/ai/book/56fa816e-042b-40dd-9a08-6bc236c079d8/page/{printed_page}","path_api_url_template":"https://workshop.closeclaw.top/api/ai/book/56fa816e-042b-40dd-9a08-6bc236c079d8/page/{printed_page}","read_url":"https://workshop.closeclaw.top/api/ai/read/56fa816e-042b-40dd-9a08-6bc236c079d8","page_url":"https://workshop.closeclaw.top/api/ai/books/56fa816e-042b-40dd-9a08-6bc236c079d8/page","search_url":"https://workshop.closeclaw.top/api/ai/books/56fa816e-042b-40dd-9a08-6bc236c079d8/search"},"page":{"printed_page":75,"pdf_page":91,"physical_index":91,"physical_index_base":0,"printed_page_status":"recognized","page_sha256":"66b0eefbfbdd51e2cb7d9713cbaaf81e834e9f51e9653a2e674b47511a08f263"},"chapter":"第三章 平等与公正：人工智能的偏见与歧视","content":"结论：人工智能并非政治中立\n\n从技术哲学中，我们知道技术在道德和政治上不是也不可能是中立的。一般来说（对于所有技术都是如此），人工智能和数据科学也是如此。虽然有些人认为，如马茨纳（Matzner 2019，109）所说的“如果人类不持续地用他们的偏见来破坏算法，算法就可以是中立的”，但这种观点是错误的。相反，人类与机器之间的关系要复杂得多，人类与人工智能也是如此。人工智能算法从来都不是中立的，社会中的偏见以及算法和数据科学过程中产生的偏见都需要进行评估。同样，正如吉特尔曼（Gitelman）和杰克逊（Jackson)(Gitelman and Jackson 2013）所言，数据本身也不是中立、客观或“原始”的。相反，它们是“由知识生产的运作产生的”，涉及解释、整理和感知—包括感觉（Kennedy，Steedman，and Jones 2020)。如前所述，人工智能使用的语料库可能存在偏见。语言本身也存在偏见，例如性别偏见。如在使用英语单词“man”时，它不仅指男性，也指整个人类（Criado Perez2019，4)。此外，正如我已经提出的，处理数据的团队也不是中立的。人们可能会有偏见，团队可能缺乏种族和性别的多样性。\n\n第三章 平等与公正：人工智能的偏见与歧视\n\n075","previous_page":74,"next_page":76}