Women in Artificial Intelligence (AI): Navigating Equity, Diversity and Inclusion (ED&I)

Synopsis

This article explores the implication of the AI for women by tackling the three issues namely: (1) Sexism and Hate Speed content generated by AI against women; (2) How to make AI more inclusive; (3) How to narrow the gap of the proportion between men and women working in the AI industry. By understanding the challenges and opportunities, we can work towards mitigating the aforementioned issues and creating a more inclusive and balanced AI landscape.

Introduction

AI technology was conceived as far back as to the 1970’s. It used to be predominantly consumed amongst the academic and research communities with a smaller portion experimented in the industry. It was not until November 2022 when the generative AI in the like of the ChatGPT by OpenAI, BingChat by Microsoft, DALL-E, Midjourney and Stable Diffusion launched in the consumer market. In March 2023, another Generative AI tool called BARD launched by Google also followed the footstep of the ChatGPT to the consumer market. Not only the IT industry but also the whole world have been taken by storm with the Generative AI technology as it has revolutionised many industries and many walks of life. Generative AI, powered by complex algorithms, machine learning and deep learning, has shown immense potential in various domains and is transforming the way we live, work, study and interact. Since the Generative AI has touched on many aspects in our day-to-day life, it is crucial to investigate its impact on ED&I, with a specific focus on women.

Generative AI – Sexism and Hate Speech against women

It has been widely reported that AI is not immune to biases ingrained in the data and algorithms upon which it is being trained. CNN has reported that AI can be racist, sexist and even creepy (source: AI can be racist, sexist and creepy. What should we do about it? | CNN Politics). The Centre for International Governance Innovation has also published an article expressing their concern that the Generative AI tools could be perpetuating harmful gender stereotypes (source: Generative AI Tools Are Perpetuating Harmful Gender Stereotypes – Centre for International Governance Innovation (cigionline.org)). The World Economic Forum has examined the rationale and logic behind why AI has a gender problem (source: This is why AI has a gender problem | World Economic Forum (weforum.org)). Social media platforms such as Twitter, Facebook, Instagram are already cracking down the prevalence of hate speech and derogatory comments posted on their sites by identifying the sources and bringing the hate speech creators to the criminal justice. However, the biggest challenge faced by these social media companies is that many of these comments are not posted by humans and instead they are generated by AI with some are even based on the algorithms generated by another level of AI. Currently, due to the lack of regulatory measures on AI and different jurisdictions in different countries, it has made it harder to prosecute the humans who create the AI algorithms. This situation has highlighted the urgent need for ethical guidelines and oversight to address bias and to ensure inclusive AI systems.

How to make AI more inclusive

Several prominent organisations such as:

have published articles covering their recommendations for gender bias. There are already case studies in place which serve as poignant examples of the impact of AL algorithms on women. For instance, the “OpenAI Case Study: Mitigating Gender Bias in AI Chatbots” (source: OpenAI’s Chatbot and the Challenges of Bias in Conversations (ts2.space)) focuses on efforts to identify and rectify gender biases in conversational AI systems. By analysing data and refining algorithms, researchers aim to ensure that AI Chatbots provide fair and unbiased responses to users, regardless of their gender. This case study highlights the importance of ongoing evaluation and improvement of AI systems to minimise gender-based biases.

Additionally, the Microsoft Research Case Study: “Promoting Inclusive Image Recognition Models” (source: Microsoft Inclusive Design) demonstrates initiatives to improve the accuracy and fairness of AI systems in recognising diverse individuals. Researchers are working to address biases in images recognition algorithms, which historically have struggled to accurately identify women and people from underrepresented groups. By training AI models with diverse and representative datasets, Microsoft aims at creating more inclusive and equitable image recognition technologies.

Skills for Women in the AI Landscape

CNN has reported that Nearly 80% of women’s jobs could be disrupted, automated by AI | CNN Business. The World Economic Forum has published a report on Why we must act now to close the digital gender gap in AI | World Economic Forum (weforum.org) and Why we need to reskill women for future of work | World Economic Forum (weforum.org). Frank Hawkins Kenan Institute of Provide Enterprise has questioned: Will Generative AI Disproportionately Affect the Jobs of Women? – Frank Hawkins Kenan Institute of Private Enterprise (unc.edu)

The above reports have demonstrated that actions are needed to support more women to actively contribute to the AI landscape. Technical proficiencies in areas such as machine learning, data analytics, algorithms, programming are required in order to develop the AI models. Beyond technical prowess, interdisciplinary skills are also vital for understanding the societal implications of AI and ensuring ethical practices.

Effective communication skills also allow women to articulate their ideas, advocate for inclusive AI design, and collaborate with multidisciplinary teams. Critical thinking and problem solving abilities enable women to identify biases and address ethnical challenges that arise in AI development. Moreover, an understanding of domain-specific knowledge also empowers women to apply AI in various industries, helping drive innovation and tackling real-world challenges.

Gender Ratio in the AI Sector

Deloitte has conducted a survey on The State of Women in AI Today | Deloitte US. The report has found that only that women make up only 26 percent of data and AI positions. Women’s representation in the AI workforce remains significantly lower than that of men which has indicated a persistent gender gap in the AI industry.

However, progress is being made through various initiatives and awareness campaigns that aim at closing this gender gap. Organisations such as Women in AI (WAI) and Women in Machine Learning (WiML) are actively working to increase the participation and representation of women in AI. These initiatives offer mentorship programmes, networking opportunities, and support networks to empower women pursuing careers in AI. By fostering collaboration and mentorship, they strive to create a more inclusive and diverse AI community.

Approaches and Initiatives to attract more women to AI

Addressing the gender disparity in AI also requires a comprehensive examination of the current educational pipeline. To encourage more young female students to pursue AI subjects, proactive measures must be taken as early as before GCSE (around 14 to 15 years of age) to allow the young female students to choose the STEM related subjects for the preparation of their GCSE and thereafter university or vocational training. .Access to coding and AI workshops can help nurture their interest and confidence in pursuing AI-related fields. Collaboration between schools, colleges, universities and industry professionals can create mentorship, internships and outreach activities that specifically target young female students, showcasing the exciting possibilities of AI.

Promoting awareness and dispelling misconceptions about AI careers is crucial. Educational institutions can play a vital role by introducing AI-related courses and programmes that emphasise inclusivity and diversity. Creating scholarship and bursaries targeted at women in AI could help alleviate financial barriers and encourage more young women to pursue AI education and training.

Moreover, mentorship programmes tailored for women in AI could provide valuable guidance and support, connecting aspiring female AI professionals with experienced mentors who can share their knowledge and skills. These mentorship programmes could help address the lack of female role models in the field and provide insights into career pathways and progression.

Collaborative efforts between educational institutions, industry leaders and advocacy groups could create a foundation to attract and retain women in AI. Partnership with companies that champion diversity and inclusion can lead to internships, job opportunities and research collaboration as these companies can provide real-world exposure and experience for aspiring women in AI. Industry conferences and events can also dedicate space to recognise the achievements and contributions of women in AI. All these efforts could lead the way to foster a sense of belonging and to inspire more young women to pursue AI careers.

Conclusion

As AI continues to shape our world, ensuring ED&I within its development and deployment is paramount. By addressing the impact of generative AI on women and fostering their contributions, we can harness the full potential of this technology while avoiding biases and creating more inclusive AI systems. Through ongoing collaboration, education and advocacy, we can build an AI landscape that reflects the diversity and talents of all individuals, empowering more women to shape the future of AI hence narrowing the gender gap in the AI industry. By embracing the talents and perspectives of women, we can unlock the full potential of AI and drive innovation that benefits society as a whole.