2025 Women in AI Expert Series: Intersectional Bias: Localization Challenges in AI and Solutions
- WAI CONTENT TEAM
- 20 hours ago
- 4 min read
Principal author: Karen Jensen
Welcome to the 2025 Expert Series from the Global Ethics and Culture office of Women in AI.
In 2025, we continue our global initiatives in Education, Entrepreneurship, Innovation, and Research to make AI accessible and inclusive for everyone, with a special focus on women and girls.
Like the 2024 Speaker Series, this year's Expert Series aims to boost opportunities for women and girls in AI. We'll feature global women experts sharing practical AI skills that could help you launch a new AI career or reskill for AI roles.
In our first session, titled "Prompt Like a Pro: AI Skills for Students and Young Professionals," our expert, Charlotte Tao, offered practical guidance on the essential skill of prompt engineering.
In our second session, titled “Agentic AI: Navigating Autonomy, Accountability, and Ethics,” our expert, Dhivya Nagasubramanian, offered clear metrics and understanding of Agentic AI and how it differs from Generative AI and its uses.
In our third session, titled, “Responsible AI in Action: A Look Back at a Winning Hackathon Project”, our expert, Dr. Ja’Nya Jenoch, brought together members of some of the teams from our Global Hackathon from 2023 and discussed where they are now, some things they’ve learned, and their inspiring message for women and girls as they move into careers in emerging technologies.
In our fourth session, titled “Generative AI, Explicit Content, and Organizational Solutions for Not Safe for Work (NSFW) images”, our expert, Bobbi Stattelman, shared concerns about the prolific generation of AI-generated explicit content and non-consensual deep fake images, and how emerging technologies are exploring organizational solutions for these challenges.
In our fifth session, titled “How AI can help Humans learn Language”, our expert, Professor Elizabeth Wonnocutt, shared insights into how Humans learn Language, the overall decline in language learning in the UK, and how her platform, Omniloquent, optimizes the language learning process.
In our sixth session, titled “Managing Hallucinations in LLMs”, our expert, Emaan Siddiqui, shares her journey into the World of AI, defines hallucinations and their negative impact on outcomes, and defines solutions to address these challenges. In this session, we explored the use of digital avatars to “interview” Emaan, so we’re excited to integrate these technologies into our work!
Our expert for today is Ishween Kuar. Ishween is a Senior Software Engineer and currently serves as a volunteer Community Lead with Women in AI California. Ishween will discuss the intersectional bias in the “English-dominant” architecture of AI platforms and tools, examining the challenges this presents for a global population of non-English language speakers, and exploring solutions to address these challenges.
THE JOURNEY
For many people around the World, their journey starts as a mix of languages. For Ishween, that start included:
· Speaking the language of her Heritage (Punjabi)
· Speaking the language of Aspiration at school (English)
· Speaking a mix of languages with her Peers (Hindi, Punjabi, and English)
From there, Ishween's journey led her to Silicon Valley, where she took a deep dive into Language Learning Models. Like one of our previous experts, Ishween quickly identified and experienced the challenges and harms of localization failures. Her personal encounter underscored a harsh reality we’ve heard repeatedly in both our Expert Series and our Global conversations: the challenge of linguistic and cultural exclusion in the tech world for women and girls.
THE PROBLEM: ENGLISH-CENTRIC AI
Despite over 63% of the world's population speaking a primary language other than English, Large Language Models (LLMs) are overwhelmingly trained on English text. This results in a one-dimensional practice that causes serious societal harm beyond simple translation failures:
Perpetuates Bias: It entrenches bias, especially against women and other underrepresented groups.
Impacts Healthcare: It can negatively affect health outcomes when non-English inputs are used in chatbots.
Hinders Employment: It puts non-native English speakers at a disadvantage in employment competitiveness.
Resource Disparity: While large companies may have the infrastructure for localization, small businesses, which represent 99% of businesses in the U.S., likely lack the necessary funding and resources to deploy this methodology. (Leppert, 2025).
THE SOLUTION: DESIGNING AI FOR INCLUSION
The core message is that equitable AI localization is not optional ethics; it's a technical necessity to prevent AI from "flattening" diverse worldviews into English abstractions. Global readiness must be engineered from day one.
The key strategies for designing for inclusion include (more in the video):
Prioritize Cultural Fit: Start with cultural fit, not literal translation.
Design with Linguistic Guardrails: Integrate real-time quality checks and flags for mistranslations, cultural tone mismatches, and toxicity in multiple localized contexts.
Measure What Matters: Track multilingual accuracy, fluency, and cultural alignment as core success metrics, alongside technical measures.
Enable Customer Choice: Give users control over their interface language and provide transparent fallback behaviors when an unsupported language is encountered, building trust.
Summary
To achieve a truly ethical AI future, we must commit to designing for inclusion from day one. This means prioritizing cultural alignment and actively engaging native speakers to ensure AI systems reflect the linguistic, cultural, and contextual richness of the world. Localization is not an afterthought; it is the pathway to building AI that serves all humanity.
Share your comments here on this post and with us on Social Media @WomeninAI to ensure we #MakeitFAIR!
Event recording: You can view the recording of the event using this link.
This Expert Series is presented by the Women in AI Ethics & Culture Office volunteer team, dedicated to A Global Vision for achieving gender parity in emerging technologies through increasing Opportunity, championing inclusive Policies, and fostering practical Action that delivers meaningful and measurable impact.
Ethics & Culture Team
Please see the links below to our Team’s profiles on LinkedIn.
Leppert, R. (2024, April 22). A look at small businesses in the U.S. PewResearch.org. Retrieved November 21, 2025, from https://www.pewresearch.org/short-reads/2024/04/22/a-look-at-small-businesses-in-the-us/