Will AI Help Break Gender Barriers—or Build Bigger Ones? | Pendo Manjele | TEDxLusaka

Key Concepts:

• Gender inequality in technology and AI
• Bias in AI algorithms and datasets
• Importance of diversity in AI development teams
• Impact of AI on women's access to opportunities (jobs, credit, healthcare)
• Need for transparency, oversight, and ethical considerations in AI
• Call to action for women, leaders, and men to address gender bias in AI

1. The Problem: Gender Disparity and Bias in AI

• Personal Experience: The speaker, a woman in technology, shares her experience of being the "only woman in the room" and feeling unheard.
• Data on Leadership Roles: A 2024 report on the state of AI in Zambia revealed that only 20% of AI-related leadership roles are held by women, compared to 79% by men.
• UN Sustainable Development Goal: Progress towards gender equality by 2030 is slipping further out of reach, according to a 2023 SDG report.
• AI's Potential Impact: AI can either be a force for gender equality or deepen existing inequalities.
• Hidden Biases: The most dangerous biases are those that silently shape our lives through AI systems.
• Real-World Examples of AI Bias:
  • Amazon's AI Recruiting Tool (2018): Penalized resumes with the word "women," favoring men.
  • Apple's AI Credit Card Algorithm (2019): Gave women lower credit limits than men, even with better financial profiles.
• Root Causes of AI Bias:
  • AI learns from biased historical data.
  • Lack of diversity in AI development teams.
  • Women make up only 30% of the global AI workforce (UNESCO report, 2025).

2. The Solution: Diverse Teams, Representative Data, and Transparency

• Need for Fairer AI: Requires more than just better code; it needs diverse teams developing, testing, and regulating AI systems.
• Representative Data: Data must reflect the full spectrum of human experience, not just segments of it.
• Transparency and Oversight: AI systems must be understood by the people they impact, avoiding "black box" algorithms.
• Women's Role: Women need to be more than consumers of AI; they need to be developers, decision-makers, regulators, and testers.
• Interdisciplinary Approach: Shaping the future of AI requires ethicists, advocates, and educators, not just technologists.

3. Call to Action: Engagement, Leadership, and Allyship

• Women's Engagement: Women, regardless of their field, need to be engaged and curious about AI's impact on their lives (healthcare, finances, jobs).
• Leaders' Responsibility: Leaders must ensure diverse teams, invest in training and mentorship for women, and assess whether their organization's tools discriminate.
• Men as Allies: Men need to speak up against bias, even before it becomes personal, and help redesign systems to be inclusive from the start.
• UNESCO Report (Ahead of International Women's Day 2025): Large language models exhibit gender stereotypes, associating women with domestic roles and men with careers.

4. Step-by-Step Processes, Methodologies, or Frameworks Explained

• Addressing AI Bias Framework:
  1. Diversify AI Teams: Ensure representation across gender, race, and background.
  2. Use Representative Data: Train AI models on datasets that accurately reflect the population.
  3. Promote Transparency: Make AI algorithms understandable and explainable.
  4. Implement Oversight Mechanisms: Establish ethical guidelines and regulatory frameworks.
  5. Encourage Engagement: Educate and empower individuals to question and shape AI development.

5. Key Arguments and Perspectives

• Argument: AI has the potential to exacerbate existing gender inequalities if left unchecked.
  • Evidence: Examples of biased AI systems (Amazon recruiting tool, Apple credit card algorithm), statistics on women's representation in AI.
• Argument: Addressing gender bias in AI requires a multi-faceted approach involving diverse teams, representative data, transparency, and ethical considerations.
  • Evidence: UNESCO report on gender stereotypes in large language models, call for interdisciplinary involvement in AI development.

6. Notable Quotes

• "The most dangerous biases are not the ones that we see. It's those that are silently shaping our lives."
• "The future that we want will not be coded by accident. It will be coded by all of us."
• "Don't just give women a seat at the table. Help redesign the table so that it was never exclusive to begin with."

7. Technical Terms and Concepts

• AI (Artificial Intelligence): The simulation of human intelligence processes by computer systems.
• Bias in AI: Systematic errors in AI algorithms that result in unfair or discriminatory outcomes.
• Large Language Models: AI models trained on vast amounts of text data to generate human-like text.
• Algorithms: A set of rules or instructions that a computer follows to solve a problem.
• Datasets: Collections of data used to train AI models.

8. Synthesis/Conclusion

The speaker argues that AI has the potential to either break down or reinforce gender barriers. The key to ensuring a more equitable future lies in actively addressing bias in AI development through diverse teams, representative data, transparency, and ethical oversight. The call to action emphasizes the importance of engagement from women, leadership from organizations, and allyship from men to shape AI in a way that benefits all of humanity. The future is not predetermined; it depends on the actions taken now to create a more inclusive and equitable AI landscape.