Is AI Biased?

Erica examines the underrepresentation of women in AI, explores the challenges in education, professional settings and societal norms, and considers the multifaceted approach needed to address why there are fewer women than men using Generative AI technologies.

Is AI Biased?

One of the most frequent questions I get asked during our keynotes and workshops is, ‘Is AI biased?’

And my answer is, yes.

Flexjobs recent report, ’The AI Gap: Exploring Variance in Workplace Adoption’, states that 54% of men are using AI in either or both their personal and professional lives, while women (35%) are adopting AI at a much slower pace.

Here are the findings broken down by gender:

Use in personal life: 15% women, 21% men
Use in both personal/professional life: 11% women, 12% men
Use at work (with manager’s permission): 5% women, 11% men
Use at work (without manager’s permission): 4% women, 10% men.

So, why are there fewer women than men using Generative AI technologies?

Well, this touches on broader issues of gender disparity in technology. This underrepresentation of women in AI needs a multifaceted approach, addressing challenges in education, professional settings, and societal norms.

According to the study by Revelio Labs and as reported by Bloomberg, AI is more likely to take over repetitive jobs, the kind that are mostly held by women.

For example, “ChatGPT can search, review, and summarise large volumes of text quickly – tasks that would normally take paralegals much longer to accomplish. In recruitment, AI can automate the process of sorting through resumes – a task that used to require more people”, Hakki Ozdenoren, Senior Economist at Revelio Labs, said.

But it doesn’t mean that other highly-skilled workers can feel secure about their jobs. Preliminary research also shows that Generative AI may impact high-wage occupations more than non-traditional manufacturing jobs.

Moving forward, “providing retraining opportunities will be key for women to navigate the evolving job landscape,” Ozdenoren said. “By doing so, we can capitalise on the potential of AI while leveraging their valuable skills and expertise”.

Michael Spencer – A.I. Writer, researcher, and curator – “Will Generative A.I. be Kind to Female Professionals? What does the data say?”

I, for one, subscribe to the school of thought that AI is not about job displacement, no matter your pronouns, age or background. AI will offer up all sorts of opportunities for careers, reskilling, and general activities in life.

Removing the bias in AI

We know there are issues in equity for women and we have much to do in this area.

Here are some of the strategies we can deploy to promote equity and fairness in our AI future and battle against these social norms:

Cultivate Early Careers
Build incredible opportunities for apprentices, interns, graduates and those reskilling and upskilling:

Integrate STEM and AI concepts into early education, using engaging and inclusive methods specifically designed to capture girls’ curiosity.
Foster welcoming STEM clubs and activities, providing young women with role models and mentors from the field.

(This one is particularly close to my heart.)

Remove Financial Obstacles
Break down any financial barriers:

Offer scholarships, grants, and financial assistance specifically for women pursuing AI and related fields, addressing economic barriers to entry.
Support initiatives that aim to make tech careers more financially accessible to women.

Build Support Systems
Connect those who want to be connected, want to learn, and can make a difference in our future:

Establish mentorship programs connecting women in AI with experienced professionals who can offer guidance and support.
Facilitate networking opportunities through conferences, workshops, and online communities dedicated to women in AI.

Champion Diversity in AI Development
Use AI to make sure that the glass ceiling gets well and truly smashed:

Promote diversity in AI research teams, recognising the value of diverse perspectives in creating equitable and inclusive AI technologies.
Implement and enforce policies that ensure equal opportunities for women in hiring, promotions, and leadership positions within tech companies and institutions.

Foster a Supportive Work Environment
Ensure everyone can see past differences, to learn from each other:

Create a workplace culture free from gender bias, harassment, and discrimination.
Implement flexible work arrangements and support initiatives that promote work-life balance for all employees, particularly benefiting women.

Advocate for Change
No longer wait around for others to do it. Those in positions of power have obligations to lead the way:

Advocate for policy changes at the governmental and institutional level to support women’s inclusion and advancement in tech fields.
Develop and enforce policies within online platforms and communities that promote safety, respect, and inclusivity, reducing barriers to participation for women.

By implementing these strategies, we can pave the way for a more equitable and diverse AI ecosystem. This fosters the creation of AI technologies that are fair, unbiased, and truly representative of the diverse world they serve.

You could argue, however, that many of these are not specific challenges to AI, but like many things, AI is merely shining a light on them. There is much for us to do, and it is all our responsibility to do it.

Read this blogpost on the TrainingZone.

References from ChatGPT v4 03/03/2024:

(2024). Discussion on enhancing participation in AI for women [Personal communication]. ChatGPT.
“Discussion on Enhancing Participation in AI for Women.” ChatGPT, 3 Mar. 2024.
2024. “Discussion on Enhancing Participation in AI for Women.” ChatGPT. March 3.