Resolving AI Gender Bias

IWD: AI Service Management is where executive AI strategy becomes operational reality. This IWD 2026, HDI Chicagoland spotlights the women shaping AI-enabled ITSM, service operations, SecOps, and AI-driven customer experience—so CIOs, CISOs, COOs, and C-level leaders can see what works, fund what scales, and lead with responsible AI governance. Join the March 11 virtual executive panel, nominate a leader, or sponsor a movement that turns visibility into performance.

Resolving AI Gender Bias is essential. With International Women’s Day around the corner, what can we do to move the dial? Women are not missing from AI. We are missing from what AI learns. When models don’t count women, outcomes in healthcare, service, and government turn inequitable. Here’s how and why to correct it.

Women Have Always Been in AI—So Why Does the Algorithm “Forget” Us?

Join us at Chicagoland HDI on March 11, to be part of the discussion, and more important the solution to when People say, “Women just aren’t in AI.” However, the truth lands harder: women are in AI—yet systems fail to count us, cite us, and surface us.

That misrepresentation changes the foundations of what AI Systems and large language models learn, what leaders fund, what conferences spotlight, and what workflows automate. Consequently, your service desk, your hospital triage, your hiring pipeline, and your benefits program inherit a male-default worldview—and then scale it.

Meanwhile, the workforce reality contradicts the myth. Women made up 47% of the U.S. workforce in 2023, even less in the past year. Exploding topics reports 72% of women report experiencing “bro culture” at work, indicating that they are experiencing gender-based discrimination and bias. 


Active keynotes and editors skew heavily male (~80% men vs ~14% women in 2024)—and those public knowledge sources influence what the internet “documents” as authority.

Additionally, OpenAI’s GPT-3 research describes training on large web corpora plus curated sources including English-language Wikipedia—meaning public web imbalance travels into models.

So when women don’t appear in the “high-authority” content stream—keynotes, conference recaps, trade publications, award lists, citations, executive bios—LLMs and ranking systems learn a distorted map of expertise. Therefore, the model under-recommends women, under-quotes women, and under-attributes women—then people call that output “neutral.”

It isn’t.

Why Conference “Keynote Exclusion” Must Be Corrected

When a conference keynote skips equitable representation of women leaders, it doesn’t just “look bad” — it actively trains the internet to keep women missing from AI leadership. That’s how visibility bias turns into AI bias in search, social, and large language models.

The visibility loop (how it happens, fast):

  • Keynotes create “receipts” the web treats as authority. Keynotes don’t end on stage; they ship transcripts, recap posts, speaker bios, backlinks, quotes, clips, and screenshots that get indexed, cited, and reused.
  • Search rewards what’s linked and repeated. Link structure and citations have long been core signals for “importance,” which is exactly why highly linked keynote artifacts travel further than quieter expertise.
  • LLMs learn from what’s most visible. Modern LLMs explicitly train on large-scale web corpora (e.g., Common Crawl) plus other sources — so the content that wins distribution becomes the content the model “knows.”
  • Existing publishing imbalance amplifies the problem. Even “neutral” knowledge sources skew: Wikimedia reports women remain underrepresented among active editors and other core contributor roles, which shapes what gets written, updated, and cited.
  • Then the model repeats the skew as “common knowledge.” People ask “top AI leaders,” the model returns a male-heavy list, that output gets reposted, and next year’s keynote planners treat the same shortlist as “validated.”

Two concrete examples of how the loop shows up in the real world:

  • Example A — “Top AI leaders” search + LLM answers
    • If your keynote lineup skews male, the web produces “Top speakers,” “Key takeaways,” and “Must-follow leaders” that skew male too.
    • Meanwhile, women remain underrepresented in AI roles overall (the pipeline is already uneven), so the default “expert set” narrows even faster.
    • Result: search results and LLM summaries converge on the same names — and they look “objective” because they’re widely cited.
  • Example B — AI in service, support, and experience
    • If keynotes ignore women leading ITSM, service operations, customer experience, and AI ethics, fewer recaps and interviews mention them.
    • That scarcity matters because models and media pull from what’s easiest to “verify” via repetition and citations.
    • Downstream, women get under-recommended for panels, advisory roles, and “expert” lists — not due to capability, but due to missing artifacts.

Best vs. worst “receipts” (credible examples you can reference):

  • Worst examples (what happens when bias ships to production):
    • Hiring AI: Amazon reportedly scrapped an internal recruiting tool after it learned patterns that penalized women’s resumes — a classic “trained on biased history → reproduces biased outcomes” failure mode.
    • Financial decisioning scrutiny: Apple Card faced high-profile allegations of gender bias; regulators later reported they did not find evidence of unlawful discrimination in underwriting — which actually strengthens the lesson: when models are opaque, trust collapses, and you still pay the reputational cost.
    • Public services + health content: Research on LLM use in local-government long-term care services found models can systematically underplay women’s health issues (among other groups), showing how “default narratives” can become operational risk when deployed for real decisions.
  • Best examples (what strong leaders do instead):
    • Build a speaker pipeline on purpose: All Raise launched Visionary Voices specifically to help organizers and media avoid “manels,” and pointed to persistent gaps in tech conference speaker representation.
    • Treat bias as a managed risk, not a PR issue: NIST’s AI Risk Management Framework pushes organizations to map, measure, and mitigate risks like harmful bias and validity failures — the same discipline keynote programming needs when it shapes “who counts” as an expert.
    • Name the problem clearly: Research cited in industry reporting shows women remain a minority of tech conference keynote speakers — meaning “we didn’t find them” is not a credible excuse; it’s a sourcing failure.

What keynote planners can do this quarter

  • Set a keynote equity bar (not just “panel diversity”) and publish it as policy.
  • Program for expertise clusters, not celebrity gravity (AI governance, responsible AI, AI + ITSM, AI + cybersecurity, agentic AI UX, human-centered AI adoption).
  • Engineer the artifacts: publish transcripts, quote cards, and recap posts that name women experts prominently and link to their work (this is how you fix “training data” in public).
  • Refuse the “single token woman” pattern: rotate formats (co-keynotes, debates, paired operator + researcher talks) so women aren’t confined to “soft topics.”
  • Audit your afterlife: track who gets quoted in recaps, whose talks get clipped, and whose names rank for “AI leader” searches 30–60 days later.

Women of AI Are Shaping the Field (Even When Conference Keynotes Miss Their Contribution)

A stage doesn’t create impact. Still, the stage signals impact—and algorithms treat signals like truth. That’s why this correction matters.

Here are women who prove the point—women already shaping AI outcomes across cybersecurity, service management, and frontier AI:

Anthropic’s leading voices (frontier AI + alignment)

  • Daniela Amodei cofounder and president of Anthropic, shaping how frontier AI companies operationalize trust, safety, and governance.
  • Amanda Askell works on finetuning and AI alignment at Anthropic, and previously worked on AI policy research at OpenAI (per her bio).

AI Women Athletes I Love to work with

  • Breanne Creelman, Global GTM Strategy Lead
    • Connects AI strategy, human-centered design, and go-to-market execution to drive real adoption.
    • Prioritizes users first experience. AI feels clear, credible, and usable—not bolted on.
    • Designs experiences that build trust, increase clarity, and sustain engagement so people adopt tools and thrive.
    • Brings a global lens; moreover, she helps teams design for diverse cultures, languages, and expectations.
  • Renate Bachmann, PMP Global AI Transformation Executive
    • Leads AI-enabled transformation across borders with compassion and strong execution.
    • Blends strategy with delivery; consequently, she builds operating models, governance, and roadmaps that turn AI potential into measurable value.
    • Accelerates global collaboration, reduces delivery friction, and scales change across diverse teams and cultures.
    • Communicates seamlessly across regions; additionally, she leverages fluency in five languages to align stakeholders faster.
  • Kim (Pham) Perez, ServiceNow Certified Technical Architect
    • Advances AI by building production-ready secure, supportable, and fair solutions.
    • Leads technical development, teaching, and mentoring stronger teams and systems.
    • Strengthens reliability and governance so AI performs consistently under real-world operational pressure.
    • Delivers practical, scalable architecture that helps organizations trust AI in live environments.

So no—women are not “missing.”
Instead, Conferences, Keynotes, Publications and AI systems keep failing to amplify the evidence of women’s leadership.

Cybersecurity + AI risk leadership

  • Rinki Sethi, CISO and CSO, Upwind Security. A Top Industry veteran security executive who has led major programs across top firms; most recently, Upwind named her Chief Security Officer, noting her decades of security leadership and prior senior roles including Twitter CISO. Importantly, cybersecurity doesn’t sit beside AI anymore—it sits inside AI. Therefore, leaders like Rinki shape how organizations defend models, protect customers, and prevent harm.

Humanising IT

ITIL Version 5 + AI governance momentum

ITIL has now moved into ITIL (Version 5), including explicit treatment of governance (and even an ITIL AI Governance exam referenced in PeopleCert’s FAQ).
Therefore, ITSM leaders—especially the women driving human-centered service, adoption, and governance—directly shape how enterprises operationalize responsible AI.

Simone Jo Moore, AI Ethicist, Thought Leader, Co-author, Senior Consultant, Master Facilitator

  • Builds community at scale through speaking, workshops, writing, and board service (including the Open Service Community), while consistently amplifying others and growing the next generation of service leaders.
  • Leads the HumanisingIT movement; therefore, service leaders treat AI as a human-tech system, not just a tool rollout.
  • Co-authored ITIL 5 and champions modern service management for AI-enabled environments, so governance, empathy, and value stay connected.
  • Shapes industry standards and mindset through recognized leadership—HDI Hall of Fame, itSMF UK Dave Jones Inspirational Leadership Award, and Thinkers360 AI Ethics & IT Strategy recognition.

Pushes practical, people-first AI adoption; consequently, teams reduce “automation without judgement,” strengthen trust, and avoid silent service failure.

HDI “top voice” leaders (service + support excellence)

HDI’s Top 25 Thought Leaders list highlights women who already influence modern service management at scale—whether or not conference keynotes prioritize them:

  • Angee Phong, Director Global IT Services- Fanatics Gaming and Betting. Recognized for consistent industry contribution and active HDI community leadership.
  • Liz Bunger, HDI national board Chairperson. Elevated for sustained advocacy and servant leadership across HDI chapters and the broader community.
  • Nancy Louisnord. Global CMO, LumApps recognized for practical ITSM guidance and visible mentorship, including support for women in tech.
  • Terri Oropeza, HDI National leader. California Bay Area Board Leader, Educator and mentor, MVP credited for years of behind-the-scenes leadership supporting local chapters and industry education.
  • Michelle Major-Goldsmith HDI Leader and Director of Service Integration. Cited for shaping global SIAM and ISO/IEC service-management standards work.

That list matters because it becomes searchable, citable, and learnable. In turn, it becomes training signal.

Other Resolving AI Gender Bias Resources

Executive Womens
Executive Womens’ Network www.linkedin.com/groups/158310/

Previus Post
IWD: AI

Leave a comment

↑