As Ege Attorneys at Law, a law firm led by women partners, we celebrate international women’s day in a novel way by addressing gender bias in AI and automation systems. We therefore prefer to draft an informative note about UNESCO’s report on Artificial Intelligence and Gender Equality (Key Findings of Global Dialogue, August 2020) . We aim at remarking significant subjects proposed by the report and thus, reckon on successful gender responsive development of AI systems hoping for advancement of non-discriminative autonomous applications.
Gender bias is one of the challenges posed by AI applications as it includes risks of having a negative impact on women’s economic empowerment and labour market opportunities caused by biased autonomous job recruiting algorithms. Moreover, unfair outcomes such as automatic vetting out of women whilst loan applications, giving women applicants lower (unfavourable) ratings or if fed by wrong data, algorithm based risk assessment in criminal justice leading to verdicts against women and more.
In its recommendation report, UNESCO opens a dialogue to discuss some key findings on gender discrimination caused by AI through machine learning based on a general overview of the current facts and experts’ contributions along with its proposal of an action plan. The report starts with an introduction with an emphasis on current threats posed by AI in terms of gender equality in particular led by AI powered recruitment systems and equal competition opportunities with men. Nevertheless, UNESCO states that the solution still lies in AI in overcoming such obstacles and has the potential to make positive changes but explicitly indicates that the real progress on gender equality must start in policy, corporate and industry decisions.
Purpose of UNESCO’s report is to identify issues, challenges, and good practices first to (i) overcome the built-in gender biases found in AI devices, data sets and algorithms; and (ii) improve the global representation of women in technical roles and in boardrooms in the technology sector; and finally, (iii) create robust and gender-inclusive AI principles, guidelines and codes of ethics within the industry. The report further provides recommendations on how to address gender equality considerations in AI principles. It also offers guidance to the public and private sectors, as well as to civil society and other stakeholders, regarding how to operationalize gender equality and AI principles.
Proposed recommendations lay some burden to stakeholders including gender advocates, practitioners, governments and academics in order to strengthen gender equality in AI principles. A collaborative work and action are demanded to ensure participation of all aiming at increasing general awareness within society at large, the gender equality community and the AI industry regarding the positive and negative implications of AI for girls, women and gender non-binary people. It is further believed that greater awareness may trigger public and stakeholder participation in AI governance. UNESCO sets other required actions such as to encourage the development of AI applications that do not perpetuate bias or negatively impact girls and women but that rather respond to their needs and experiences; create funding mechanisms for participatory AI, access to AI and AI education for girls and women; promote diversity and equality through hiring, supply chain and related practices; and contribute to the collection of sex-disaggregated data and so.
In the light of the UNESCO’S report and existing facts, if AI and automation are not developed and applied in a gender-responsive way, it seems that they are likely to reproduce and reinforce gender stereotypes and discriminatory social norms. Therefore, it is crucial that stakeholders (i.e. governments, companies, software designers and other actors) take concrete steps to achieve goals and works set by the report but most importantly more women must actively participate in creating AI and automation systems rather being sole beneficiaries.
For more and detailed information please contact firstname.lastname@example.org or email@example.com