Workplace Diversity Metrics

"The curriculum was never neutral—it was just never built for us." Black children walk into classrooms and immediately begin the work of code-switching, translating, suppressing, and surviving. White children walk in and see themselves— in the books, in the language, in the heroes, in the holidays, in the classroom norms. They are not just learning content—they are absorbing validation. What Happens in the Brain? The brain latches onto familiarity to process new information. That means: Relevance = retention Cultural connection = cognitive ease Representation = emotional safety Without these, cognitive load skyrockets—and learning stalls. Culture Fuels Comprehension Research shows that culturally responsive pedagogy: Increases reading comprehension Activates prior knowledge more efficiently Enhances motivation and attention Supports metacognition and critical thinking The Real Problem? We keep trying to close the reading gap with instruction, while ignoring the cultural gap baked into the curriculum. “You cannot separate literacy from identity. If students don’t see themselves, they don’t see the point.” —Dr. Gwendolyn Battle Lavert

Not all representation is created equal. For example: 1. Showcasing faces of various identity groups in advertisements but lacking diversity in executive leadership or decision-making roles. 2. Hiring someone from a marginalized background but failing to provide mentorship, resources, or a safe work environment, eventually setting them up for failure. 3. Promoting a leader from a marginalized group, but that leader lacks real influence or is pressured to conform to the status quo rather than advocate for systemic change. 4. Celebrating cultural holidays but refusing to address pay gaps, biased hiring practices, or harmful behavior. Oftentimes, leaders believe people will overlook these performative gestures, but people are well aware of the intention behind them. —— [Image description] Text on a beige background that reads: Representation without access, power, and protection is merely tokenism and virtue signaling. It creates the illusion of progress while perpetuating existing systemic barriers and inequities.

They might *look* the same in your company photo, but let's be clear: 𝗗𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆 ≠ 𝗥𝗲𝗽𝗿𝗲𝘀𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 ≠ 𝗧𝗼𝗸𝗲𝗻𝗶𝘀𝗺... and it matters. Because while all three might show up in your company at some point, only ONE is going to actually move your company culture forward - and there are FOUR things you can do to help that forward motion happen today. So, let's break it down: 𝗗𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆 is the presence of meaningful difference. It’s when our companies have people in it with various experiences and demographics and realities. This is *potentially* great, but it’s potential - not always or evenly realized. 𝗥𝗲𝗽𝗿𝗲𝘀𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 is when people bring both presence and perspective. It’s when a person from a particular background is not only present, but also connected to and invested in the communities they represent. They bring welcomed insight, advocacy, and lived understanding along with their differences. That’s a𝘱𝘱𝘭𝘪𝘦𝘥 diversity. 𝗧𝗼𝗸𝗲𝗻𝗶𝘀𝗺, meanwhile, is what happens when presence is mistaken for progress. It’s when someone checks demographic boxes on paper, but their affiliations with those groups may not be real, relevant, or respected. In any case, they’re not given the power, trust, or support to shape decisions with any insight or perspectives informed by their affiliations. That’s 𝘥𝘦𝘤𝘰𝘳𝘢𝘵𝘪𝘷𝘦 diversity. In short: Diversity = Presence of Differences. Representation = Presence + Perspective. Tokenism = Presence – Power. So, people… let’s aim for representation this week, eh? And here’s what we can do today to make that happen: 💡 1. Shift from “having” to “hearing.” Instead of asking “Do we have diversity?” ask “Whose perspective have we actually considered in this decision?” (It’s a small change in language that reframes presence into power.) 💡 2. Listen with intent. Make space for someone whose perspective you’ve heard less of - and ask what would make them feel more included or supported this week. 💡 3. Share the mic. If you’re in a meeting or planning something, invite someone knowledgeable from a different background to co-present, co-author, or co-lead. Because, ya know... diversity grows into representation when power and credit (not just visibility) are shared. 💡 4. Audit the diversity you celebrate. Look at one area - an employer branding campaign, event, or team photo - and ask: is this representation or decoration? If the people featured weren’t part of the process or the power - if they lacked a say in their own stories - then that's a sign to re-engage more meaningfully. So, yes…. There’s *so* much we can do, so many ways to do better today... and, as always, I'm here cheering you on so we can do better together. #Representation #Inclusion #Diversity #DEI #Leadership #WorkCulture #WorkUPWARD #EmployerBranding

Another academic paper I've loved. This one looks at 125,000 movie posters from over 35,000 films to quantify how actors of different ethnic backgrounds are positioned, sized and selected for film posters. Key takeaways: 1. White actors still dominate poster design Despite changes on screen, poster visibility lags behind. Until recently, White actors made up over 79% of the faces on posters, far exceeding their proportion of the US population. In 3-person cast posters, only 9.2% of top-billed faces were non-White. 2. Ethnicity affects face size and position White actors were consistently larger and more central on posters. Minority actors, when included, were more likely to be smaller and positioned away from the centre (visual cues that influence perceived importance). 3. Genre stereotypes persist Minorities appear more often in specific genres. Black actors are more visible in crime films, Asian actors in action. Representation in Westerns, Mysteries and Film-Noir remains almost entirely White. 4. Poster diversity is improving… but just recently Posters from the last two years now reflect the ethnic makeup of the US population. This suggests a shift in design priorities, perhaps due to industry pressure, franchise casting or global market strategies. 5. Posters still echo casting hierarchies White actors appear disproportionately in top-ranked roles, while Black and Asian actors are more present in supporting cast positions. The visual hierarchy on posters mirrors these ranks. The full paper is here: https://lnkd.in/eHn2EePp By Dima Kagan, Mor Levy, Michael Fire & Galit Fuhrmann Alpert Kagan, D., Levy, M., Fire, M. et al. Ethnic representation analysis of commercial movie posters. Humanit Soc Sci Commun 11, 180 (2024). https://lnkd.in/eHn2EePp 

📢 #NewPaperAlert 🚨 Thrilled to share our latest research on LLM Alignment & Cultural Representativeness, “Mind the Gap: Pitfalls of LLM Alignment with Asian Public Opinion” supported by EkStep Foundation. 🔍 What we studied Audited alignment of leading LLMs (GPT-4o, Gemini-2.5, Llama 3.2, Mistral, Gemma 3) across India, East Asia, and Southeast Asia Focused on the sensitive and high-impact domain of religion as a lens to evaluate broader cultural alignment Evaluated across multiple languages, not just English ✅ Key Contributions Introduced a multilingual representativeness audit framework, comparing model opinion distributions against nationally representative Pew survey data Benchmarked with CrowS-Pairs, IndiBias, ThaiCLI, and KoBBQ for culturally grounded bias testing Proposed lightweight steering interventions (demographic priming, native-language prompting) to mitigate gaps 📊 Key Findings Models align well on broad social issues, but systematically misrepresent religious viewpoints, especially for minority groups Even in local languages, misalignment persists, suggesting deep training-data biases Small interventions can reduce—but not eliminate—cultural gaps 🌏 Why this matters LLMs are being deployed globally, yet remain trained on English-centric corpora. Without regionally grounded audits, we risk systemic under-representation and reinforcement of harmful stereotypes. Work w/ Hari Shankar, Vedanta S P, Sriharini Margapuri, Debjani Mazumder, Abhijnan Chakraborty 💻 🧵 📄 Code & Data, X thread, SARAL Video and Preprint in comments #LLM #ResponsibleAI #CulturalAlignment #AIethics #BiasAuditing #AIResearch /c Precog-at-IIITH

Every time a leader steps onto foreign soil, they carry more than a passport and a policy brief. They carry a story. A people. A culture. President John Mahama’s recent moment with the fugu (batakari) is a powerful reminder of this truth. What began as a simple choice of attire during a visit to Zambia sparked global conversation, mislabelled by some, debated by many, and ultimately spotlighting Ghanaian culture on an international stage. What some might have dismissed as a trivial fashion moment became something deeper: cultural diplomacy in action. Culture travels faster than speeches. It reaches hearts before policies do. When leaders proudly wear, speak, and embody their cultural heritage abroad, they do more than make a statement, they create visibility. That visibility can translate into economic opportunity, especially for artisans, creatives, and local industries whose work often remains unseen beyond borders. In this case, smock weavers and designers were suddenly part of a global conversation. That is not accidental impact; that is leadership with cultural intention. Imagine if more leaders understood this power. A fabric can become an export. A tradition can become a global brand. A misunderstood moment can become a gateway to pride, jobs, and soft power. In a globalized world where identities often blur, authentic cultural representation is not a step backward, it is a strategic step forward. Leaders don’t just represent nations through agreements and diplomacy; they represent them through symbols, stories, and everyday choices. So here’s the call: When leaders travel, let them travel rooted. Let them wear home. Speak home. Carry home with confidence. Because when culture is embraced at the highest level, it doesn’t just stay cultural, it becomes economic, diplomatic, and transformational. Culture is not decoration. It is power. Picture credit: DW #Africa #culture #leaders #leadership #LinkedIn

When AI Reinvents Culture — And Why We Must Correct It Artificial intelligence is expanding faster than our cultural imagination. It doesn’t arrive as perfect neutral code. It arrives carrying the imprints of the data it was trained on — data shaped by particular histories, perspectives, and biases. This is why generative models have been caught reproducing stereotypes across languages. Margaret Mitchell, Chief Ethics Scientist at Hugging Face, showed in the SHADES dataset how models trained largely on Western material not only echo biases, but amplify them and spread them into cultures where those distortions never existed. What begins as an imbalance in English becomes a global distortion. The problem isn’t abstract. Google’s Gemini system recently generated images that placed Black individuals in Nazi-era scenes. What was meant as a correction for representation turned into an overcompensation that produced something worse: a collapse of historical context. A single misaligned adjustment revealed how fragile cultural nuance can be inside a system that doesn’t truly understand it. Yet research also points to a way forward. Cornell University demonstrated that a simple instruction — a cultural prompt telling the model to respond as if it were from a specific region — reduces bias across more than one hundred countries. A single sentence shifts the worldview of the machine, altering not only the answers it gives, but the lens through which it perceives the question. The lesson here is clear. AI is not culture-agnostic. It carries the shadows of its data. When these shadows are left unchecked, they grow larger, distorting meaning, undermining trust, and turning diversity into caricature. But it is also clear that cultural tuning is possible. With deliberate design, AI can learn to sense difference instead of flattening it, to honor nuance instead of erasing it. If AI is to serve globally, it must learn to speak culturally. Otherwise it does not bridge divides. It multiplies them.

Whose morality is AI learning? And whose values is it quietly amplifying? A recent study published in the Proceedings of the National Academy of Sciences (PNAS) raises a critical question for the future of Artificial Intelligence: Can AI truly serve humanity if it misunderstands humanity’s cultural diversity? Researchers compared the moral judgments of GPT models with responses from more than 90,000 people across 48 countries. The findings are revealing. Even when AI was instructed to respond as an "average citizen" of different nations, it consistently leaned toward moral priorities commonly associated with Western societies—such as individual rights, care, and equality—while underrepresenting values that many non-Western societies place greater emphasis on, including purity, tradition, loyalty, and respect for authority. This is not merely a technical issue. It is a governance issue. As AI becomes embedded in education, journalism, healthcare, public policy, workplace decision-making, and digital platforms, the values encoded within these systems will increasingly shape how information is presented, what content is promoted or moderated, and how advice is delivered. The challenge is not whether Western values are right or wrong. The challenge is whether AI systems can accurately reflect the diversity of human values rather than unconsciously exporting a single cultural worldview as a universal norm. For countries in Asia, Africa, the Middle East, and Latin America, this raises important questions: 🔹️Who defines the ethical foundations of AI? 🔹️Whose cultural perspectives are represented in training data? 🔹️Can AI governance frameworks ensure cultural pluralism rather than digital homogenization? 🔹️Are we building globally inclusive AI, or digitally reproducing existing power imbalances? As governments around the world develop AI policies and regulatory frameworks, cultural representation must become a core component of AI governance, fairness, and accountability. The next frontier of responsible AI is not only reducing algorithmic bias—it is ensuring that AI understands and respects the rich diversity of human values, traditions, and social norms. Because an AI system that speaks every language but understands only one worldview cannot truly serve a global society. Digital Cooperation Organization (DCO) Justice and Universal Rights Initiative UNESCO United Nations United Nations Population Fund (UNFPA) Digital Rights Foundation Global Engineering Network for Innovation and Intelligence (GEN_II) Prime Minister's Youth Programme Youth Impact Netwrok(YIN) YES! YOU CAN Global Foundation since 2018 Institute of Administrative Sciences, University of the Punjab, Lahore - Pakistan #ArtificialIntelligence #AIGovernance #ResponsibleAI #EthicalAI #AIPolicy #DigitalGovernance #CulturalDiversity #AIEthics #TechnologyPolicy #InclusiveAI #DigitalTransformation #PublicPolicy #GlobalGovernance #EmergingTechnology #FutureOfAI

The first time I cried in a movie theater was when I saw "The Joy Luck Club". My high school Asian club went to see the movie as a group, and I remember so clearly that, for the first time, I saw not only a story that I could relate to, but a story that I could relate to because I saw *myself* in that story, as an Asian American woman. That's why I'm always going to fight so hard for representation. It matters. Today's news highlights just that (especially for people who are waiting to hear "the business case" for diversity, equity and inclusion). Because, beginning in 2021, the consulting firm McKinsey has published a series of reports exploring the entertainment industry’s representation and inclusion of historically excluded people. Along the way, the analysts totaled up the potential financial revenue that Hollywood could stand to gain if it adopted more culturally inclusive business solutions: $10 billion per year from closing the Black inequity gap, an eye-popping $12 billion to $18 billion from properly valuing Latino professionals and consumers, and — in the latest report released today — $2 billion to $4.4 billion from more effectively tapping the Asian and Pacific Islander market. That's $30 billion dollars being left on the table. Some parts from today's report that stood out to me, thanks to the Hollywood Reporter article covering this, include: 💡 Although API representation in film leapt from 3 to 20 percent in the 20 years between 2002 and 2022, 85 percent of that recent representation is in movies produced outside the United States, and episodic television has seen similar trends. AAPIs (6.2 percent of the U.S. population) are still underrepresented by 50 percent in content made in their own country (3.4 percent of lead roles in U.S.-produced films), which means that the API faces seen onscreen do not necessarily reflect the experiences and perspectives of Asian Americans and Pacific Islanders living here. 💡 Nearly half of wide release features with API leads are action/adventure movies (for films that grossed more than $50 million, that figure rises to 71 percent), and movies with API leads tend to be increasingly race-agnostic (as opposed to having narratives specific to an API cultural experience) the more widely they are distributed. 💡Disaggregating representation data – by genre, by producing country, by qualitative analysis – is particularly crucial for the umbrella identity known as API, which tends to obscure or outright erase the Pasifika (Native Hawaiian and Pacific Islander) part of the moniker. Half of the 310 API lead roles in movies from 2018 and 2022 were played by actors of East Asian descent; the 17 Pasifika leads in those five years were played by just five men, who were mostly in action movies and who were mostly Dwayne Johnson and Jason Momoa. Full article and report linked in comments. Would love to hear your thoughts on this as well. #AANHPI #representation #matters

96% of our knowledge of humans comes from 12% of the world's cultural contexts. Our understanding of humans needs to stop being so WEIRD. Western. Educated. Industrialized. Rich. Democratic. The study of human development specifically has a representation crisis: A 2023 study found that 84% of studies relied exclusively on data from geographical regions inhabited by less than 7% of the world’s population. This is more than just bad science. These biases in our research directly shape how we define “good” babies/parents: ⇢ Attachment measures standardized on Western families pathologize cultural caregiving practices around the world ⇢ Developmental assessments fail to account for cultural variation in milestones ⇢ Mental health screening tools assume nuclear families and individualistic values Here's your action item. First, ask yourself: Leaders: Do your policies account for diverse family structures and ways of learning? Maternal care providers: How are your assessment tools validated across cultures? Policymakers: Whose development is centered in your early childhood initiatives? Researchers: Who designs your research questions and methods? Now, apply 5 ways to combat these biases: 1. Re-examine your measurement tools - Validate assessments across cultures - let cultural experts lead tool development - Question Western-centric assumptions 2. Challenge your frameworks - Question "universal" theories - Consider cultural epistemologies - Recognize multiple pathways 3. Transform your training - Include cultural humility education - Center diverse family structures - Challenge deficit-based perspectives 4. Adjust your implementation - Allow flexible delivery methods - Adapt to local caregiving practices - Support indigenous knowledge systems 5. Share power - Let communities lead research design - Support community-led initiatives - Redistribute research resources Remember: Good science requires representation. Better representation creates better science. ↓ What’s an example from your work that made you question whether an approach was actually inclusive? How have you overcome cultural biases? - • • • • • • • • • • • • • • • • • • • • • • • • • • • For more on research + babies ↓ Follow Emily Little, PhD Join my newsletter: https://lnkd.in/gCJa6pM5