AI is entangled in every aspect of our lives, from the workplace to healthcare. Trillions are being invested in the promise that it will make things faster, smarter and more accurate. But artificial intelligence can only ever be as good as the data we feed it — and thanks to the gender data gap, most AI models have been trained on skewed data which leads to skewed decision-making, missed opportunities and poorer outcomes. But it doesn’t have to be this way. In this talk Caroline lays out the possibilities and pitfalls of this new tool, identifying key lessons from the past and crucial steps for the future that will enable us to harness the promise of AI for everyone.
What Caroline Explores
Why is it important to address data bias as we build the future of AI? In this talk, Caroline examines how:
- AI will only be as effective as the data it is trained on
- AI learns patterns from data. If that data underrepresents women, mislabels gender, or reflects societal stereotypes, then the AI model will learn and amplify those biases. This is why:
- Hiring algorithms trained on past data where men were hired more often for tech jobs can learn to rank women lower.
- Voice assistants may recognize male voices more accurately because male speech data was overrepresented.
- Health prediction models may miss conditions that manifest differently in women (e.g., heart disease or pain response).
- AI learns patterns from data. If that data underrepresents women, mislabels gender, or reflects societal stereotypes, then the AI model will learn and amplify those biases. This is why:
- AI-driven tools can either widen or close the gender data gap
- AI tools reflect the biases of the data sets that are trained on. To ensure these tools benefit all people, they need to address the gender data gap and make sure that they are not merely amplifying existing issues.
- Algorithms and analytics built on biased or incomplete data risk reinforcing inequities rather than fixing them.
- A 2025 study titled “Who Gets the Callback? Generative AI and Gender Bias” audited open-source LLMs across ~332,000 real job postings. The study found that women are less likely to be recommended (i.e., receive “callbacks”) than men especially for higher-wage, male-dominated occupations.
- A 2025 study found that AI tools used by English councils downplay women’s health issues, perpetuate regressive gender stereotypes
- But when AI reflects accurate data it can help diagnose heart attacks in women, radiologists spot breast cancer in real-world tests and design safer seat belts.
Why Book Caroline Now
We are living in a time of rapid digital transformation, AI expansion, global economic uncertainty, and a push for social equity — all of which make inclusive, accurate data more critical than ever.
Caroline’s keynote addresses the need for data-driven campaigners and shows how to make meaningful change in your own community.
Learn how to spot and correct biased data:
- As AI tools and machine learning systems are adopted more widely, they rely heavily on historical data — much of which underrepresents or misrepresents women.
- Organisations using biased data to train AI risk baking inequality into algorithms, whether in hiring, healthcare, finance, or customer service.
- Fixing the gender data gap ensures AI works fairly for everyone.
Design resilient and inclusive systems:
- Climate change, economic instability, and conflicts disproportionately affect women, but gender-disaggregated data is often missing in disaster response, policy planning, and resource allocation.
Build workplaces that foster talent and deliver sustainable results
- Despite progress, pay gaps, underrepresentation in leadership, and barriers in tech and STEM remain. Gender-specific data on recruitment, retention, promotion, and pay is necessary for effective action. And, in 2025, investors, regulators, and employees are demanding data-driven transparency around gender equality.
- In 2025, markets are saturated — but underserved groups (including women) represent untapped opportunities. Accurate gender data allows organizations to design products, services, and policies that will answer real needs and win customer loyalty.
Why Audiences Choose Caroline
Caroline’s presentations have fundamentally shifted how scholars, policymakers, and practitioners think about data, evidence, and knowledge creation. Her speaking engagements consistently receive praise for making complex research accessible while maintaining intellectual rigor.
Through her keynotes and presentations, Caroline helps audiences understand that inclusive approaches to data are not merely matters of fairness, but essential components of accurate understanding and effective action.
Her ongoing commitment to rigorous inquiry and engaging communication continues to illuminate new dimensions of how knowledge is constructed and how it might be made more complete.
What Audiences Are Saying
It was a fantastic event. The talk itself could have continued for another hour!eBay
An eye opening presentation and I know that our employees gained a lot from taking the time and listening in on it. J.P. Morgan
Caroline was fantastic and the feedback from our team has been glowing!…It was an eye opening conversation for so many on our team.
MIO Partners
We absolutely loved hosting Caroline. She was superb, engaging and left our audience with the facts and data they need to address gender bias in both their personal and professional lives.
Trouble Club
Speaking Engagements
Caroline accepts a limited number of keynote presentations each year, focusing on opportunities to address leading organizations and institutions who are working to make the world a better place.
Each presentation is thoughtfully prepared for the specific audience and context, ensuring relevance and depth appropriate to the questions being explored.
To enquire about booking Caroline please contact Michael at [email protected]