From Epistemic Injustice to AI Ethics: A Frickerian Approach to Gender Discrimination

Testimonial Injustice occurs when the speaker is wrongfully undermined in his/her capacity as a knower due to the identity prejudice against his/her, while Hermeneutical Injustice occurs when some social groups cannot make sense of their social experiences due to their exclusion from hermeneutical participation. Both kinds of injustice result from structural social prejudice and can especially disadvantage some groups but not others. In the real world, these two types of injustice are embodied in the design and practice of AI. Based on gender-biased data, AI further exacerbates the Testimonial Injustice women suffered when processing and accessing information. And the low participation rate of women in occupations such as Science, Technology, Engineering, Mathematics results in the low diversity and inclusiveness of AI technologies and exacerbates the Hermeneutic Injustice faced by women. Therefore, in the era of AI, these two types of injustice are practiced by new technologies and manifested in the aggravation of existing gender biases against women.