Most people feel the effects of artificial intelligence long before they understand how it works. They notice it in a LinkedIn debate that flattens gender into one storyline, or when an image generator subtly shifts their features toward a stereotype.
In more harmful cases, communities already facing discrimination experience AI through increased surveillance and policing. These moments raise real questions about safety and autonomy, even when people can’t point to the exact technology shaping their experience.
What makes this confusing is how little anyone is allowed to see. The data, the modeling choices, the rules all sit behind intellectual property protections. People are expected to trust tools they cannot review, while algorithms read their faces and decisions out of sight.
What is visible is the history inside the data. These systems were trained on information shaped by decades of inequality, and that history shows up in their outputs.
PURPOSE OF THIS GUIDE
In my work across strategy, digital transformation, and AI change management, I keep seeing conversations that never reach the real issue: POWER. When that force goes unexamined, people focus on symptoms instead of structures.
In the pages ahead, I walk through power, identity, and digital autonomy, and how each influences the way AI touches people and communities. I also introduce the EDEN framework, a model for engaging digital systems with ethical awareness and responsibility.
My goal is simple: To give you the clarity and language to protect your voice, challenge harmful outcomes, and participate in public conversations with real agency.
Power determines what becomes possible in a society. It decides whose experiences are treated as credible, whose concerns are ignored, and whose worldview becomes the basis for public policy, business strategy, and now, artificial intelligence. AI is rarely introduced as a political actor, yet every model carries the imprint of the social, economic, and institutional forces that shaped it.
HOW AI EXPANDS POWER
Once deployed, AI systems scale the assumptions they were built on. A biased dataset or flawed rule, once embedded into a school, workplace, hospital, or bank, can influence thousands of decisions. What institutions believe is neutral technology often reinforces old hierarchies, turning individual acts of discrimination into something widespread and automated.
This imbalance deepens when we consider who can actually challenge these systems. The people most affected by AI harm rarely have the ability to inspect the tools shaping their opportunities. Companies protect their models as proprietary, agencies cite national security, and workers, students, patients, and community members are told to trust automated classifications they cannot examine.
Beyond individual decisions, AI also shapes what the public comes to believe is true. Systems that filter and recommend content now influence what appears trustworthy and whose expertise is taken seriously. This is epistemic power: the ability to define what counts as knowledge. When posts about bias written by Black or trans women lose visibility while similar messages from white cisgender women get elevated as thought leadership, the public conversation narrows without anyone making an official announcement.
The pattern becomes even clearer at the global level. Many AI models rely on data sourced from countries and communities with fewer structural protections and less influence over how the technology is developed. Places across Africa, South Asia, Latin America, and the Middle East often contribute data without receiving equal access, benefits, or representation in the decision-making process. When governments with limited leverage depend on systems built and controlled by foreign corporations, they inherit the values, assumptions, and priorities of institutions that aren’t accountable to them.
In conversations about AI, identity is often flattened into single labels like “woman,” “Black,” or “immigrant,” as if each category exists in isolation. That simplification obscures how systems actually distribute opportunity and harm.
Intersectionality offers a more accurate way to understand identity inside systems of power. Kimberlé Crenshaw coined the term in 1989 while analyzing discrimination cases where Black women sued their employers. The courts dismissed their claims because the legal system treated “race discrimination” and “gender discrimination” as separate issues. Crenshaw showed that Black women faced forms of discrimination that could not be understood by isolating one identity from another.
Intersectionality then emerged as a framework to explain how identities overlap to create distinct vulnerabilities within structures never designed to recognize or protect them.
WHY INTERSECTIONALITY MATTERS WITHIN AI
Most technological systems replicate the same structural gaps Crenshaw identified in the law. Algorithms treat identity categories as discrete variables, if they recognize them at all. They don’t account for how race, gender, class, disability, sexuality, immigration status, or religion interact. Instead, they rely on simplified categories that assume a default user: white, cisgender, male, able-bodied, and economically secure.
When AI systems ignore how identities overlap, the failures follow predictable patterns:
- Facial recognition systems misclassify darker-skinned women at higher rates because the system’s design and training data centered lighter-skinned men.
- Healthcare algorithms omit medication recommendations for an attention-deficit/hyperactivity disorder case when race was explicitly stated, but suggest them when those characteristics were missing from the case.
- Image generators sexualize Asian women because the training data reproduces stereotypes embedded in media and popular culture.
Identity also shapes risk. Immigrant communities now face algorithmic surveillance not only at borders but throughout US cities, where ICE relies on data brokers, facial recognition, and predictive tools to flag “risk” or locate individuals. Disabled people are misjudged by automated eligibility systems in healthcare and social services. Muslim communities continue to be profiled by predictive security systems built on decades of political bias. These patterns show how identity collides with institutional priorities long before individual behavior even enters the picture.
AI doesn’t simply use data. It creates meaning from it. Every search, purchase, location signal, and scraped image becomes part of a shadow profile that circulates far beyond where it was gathered. Data brokers assemble these fragments into detailed portraits, then sell them to advertisers, insurers, political campaigns, and employers.
Much of this happens without consent. Even if you opt out of social media, the ecosystem still learns from the people around you, inferring facts about your identity with surprising accuracy.
This is why privacy alone no longer protects people. Race, income, religion, and health status can be estimated without a single disclosure. Once made, those inferences quietly influence hiring decisions, credit limits, border checks, and the content someone sees online. The person at the center of these decisions may never know a profile exists, much less how it defines them.
For organizations, autonomy breaks down the moment internal data flows into AI systems they cannot audit or govern. Once that happens, teams are left accountable for decisions they can’t fully explain, even when those decisions carry real consequences for employees or customers. This loss of visibility creates real anxiety for leaders, who know that a single flawed automated decision can expose the organization to discrimination claims, damage public trust, and draw scrutiny from regulators demanding explanations they may not be able to provide.
HOW AI PLATFORMS SHAPE BEHAVIOR
Loss of digital autonomy doesn’t only happen at the system level. It also shows up in daily behavior. Social media platforms teach users how to behave, rewarding certain tones or narratives while burying others. Over time, professionals reshape their voice to match what the feed appears to prefer. They adjust their language, avoid certain topics, or simplify complex ideas to avoid being penalized.
This affects everyone differently. Executives avoid criticism that might threaten their position. Marginalized groups self-edit for physical and mental safety. Some adapt deliberately. Others absorb the cues unconsciously. But the pattern is consistent: when people tailor themselves to algorithmic expectations, the system defines the boundaries of public discourse.
Once the connections between power, identity, and autonomy become clear, the question becomes more practical:
How do we engage in digital spaces in ways that reduce harm and protect our agency?
That question is the starting point for the EDEN framework.
EDEN stands for Ethical Digital Engagement Norms. It’s a four-pillar framework that helps people evaluate the social forces, technical design choices, identity dynamics, and behavioral patterns shaping digital experiences. The framework guides decision-making before posting, speaking, designing, or leading inside AI-shaped spaces.
EDEN’S FOUNDATION: INTENT VS IMPACT
The foundation of EDEN comes from a principle I teach within my Inclusive Coaching courses for facilitators, executives, and coaches:
Intent should NEVER eclipse impact.
Good intent cannot erase harm or repair a relationship. What matters is the effect of our choices. This is especially true in digital environments, where platforms amplify and distort messages in ways no one can fully control. What someone shares in frustration or excitement can take on a life far beyond the original meaning.
EDEN is not a strict rulebook. Its purpose is to support reflection, accountability, and intentional action. It provides language and structure to help people pause before acting, understand the forces shaping a conversation, and engage in a manner that matches the impact they wish to have.
HOW TO USE THE EDEN FRAMEWORK
EDEN is built on four core types of awareness. Each one addresses a different dimension of how power, identity, and technology interact in digital spaces:
- Social Engagement
- Technical Engagement
- Intersectional Engagement
- Behavioral Engagement
Each pillar breaks down into two parts: ethical norms and questions. The norms show you what responsible engagement looks like in that area. The questions help you apply those norms to your specific situation.
Whether you’re building an AI system, navigating a digital platform, or making decisions that affect how people experience technology, start by reading through the pillar that fits your situation. Then work through the questions before you act.
Some moments will require reflecting on all four pillars, while others need just one. The key is slowing down long enough to understand the forces shaping your choices.
1. SOCIAL Engagement
Conversations build, collide, and shift long before you decide to join in. Pay attention to the history people bring to a topic, the tension beneath it, and the communities already shaped by it, so you don’t enter blindly or center yourself.
ETHICAL NORMS FOR ENGAGEMENT
- Account for unequal access to safety: Acknowledge that different groups experience digital risk differently. What feels harmless to you may land dangerously for communities who have been surveilled or targeted.
- Avoid crisis opportunism: Resist the urge to post, comment, or “take a stand” in moments of public tension when you do not understand the full context or the people most affected.
- Notice who is shaping the narrative: Identify whose voices are currently centered, whose perspectives are missing, and how the platform’s algorithm widen or narrow public understanding.
- Pause before adding your interpretation: Consider whether your contribution brings clarity or shifts attention away from the communities with lived experience.
QUESTIONS TO ASK BEFORE ACTING
- Which communities have already been affected by this issue, and how have they described their experience?
- How might this message land for people whose social or historical context is different from mine?
- What events, tensions, or power dynamics are shaping how this topic is being interpreted right now?
- Whose voices have shaped the narrative so far, and whose perspectives are missing or minimized?
- Does my contribution clarify, complicate, or distract from what the most affected communities are saying?
2. TECHNICAL Engagement
Every digital tool is built on decisions most people never see. Look underneath the interface to understand how a system was shaped, where it will fail, and whose values drove its design.
ETHICAL NORMS FOR ENGAGEMENT
- Know the limits of the system: Understand what the model can and cannot detect, classify, or interpret before relying on it.
- Trace the data’s origin: Examine where training data comes from, whose histories it reflects, and which communities are missing from it.
- Document design decisions: Make clear how choices were made in training, filtering, modeling, and evaluation, and who those choices benefit or disadvantage.
- Anticipate failure modes: Expect errors, bias, and blind spots, especially in contexts involving risk, identity, or uneven access to safety.
- Prioritize transparency where possible: Provide context about limitations and uncertainty so people understand how the system arrived at its interpretation.
QUESTIONS TO ASK BEFORE ACTING
- Which identities or communities are missing from the training set, and how might that shape the model’s behavior?
- What definitions of accuracy, quality, or safety were used during development?
- Who benefits from the way this system is designed, and who is placed at risk?
- What are the known failure patterns of similar tools?
- If this tool makes a mistake, how visible will that harm be, and who bears the consequences?
- What information or transparency is needed for people to use this technology responsibly?
3. INTERSECTIONAL Engagement
Digital harm hits hardest when identities intersect in ways systems weren’t designed to recognize. This lens keeps the people at highest risk at the center of how you build and engage.
ETHICAL NORMS FOR ENGAGEMENT
- Reject the default user: Assume digital systems were not designed with everyone in mind, and do not treat one identity label as the stand-in for an entire community.
- Honor lived experience as expertise: Treat people with lived experience as primary sources, and do not replace what affected communities say about themselves with your own interpretation, even when you mean well.
- Name what the system cannot see: When a community is erased, misclassified, or underrepresented by a tool or platform, say so clearly and treat low visibility as a sign of exclusion, not disengagement.
- Design and communicate with nuance: Instead of simplifying identities to make things easier for you or the algorithm, leave room for complexity.
QUESTIONS TO ASK BEFORE ACTING
- Who is placed at the highest risk if my engagement is misunderstood or misclassified?
- What identities or combinations of identities are most likely to be excluded or overpoliced?
- Does this message, tool, or policy assume a single “universal user”?
- How might this land for someone whose lived experience includes multiple intersecting forms of marginalization?
- Whose lived experience is shaping this decision? Whose is missing?
4. BEHAVIORAL Engagement
Technology doesn’t post for you. You decide how you show up. Own the tone you set, the impact you create, and the repair you owe when harm occurs. In digital spaces, intention never outweighs impact.
ETHICAL NORMS FOR ENGAGEMENT
- Acknowledge harm without defensiveness: When someone names harm, respond with curiosity and responsibility rather than centering your own feelings.
- Practice repair, not justification: Focus on what would rebuild trust or reduce harm, not on proving you meant well.
- Pause before responding: Give yourself time to understand the landscape, your role in it, and the potential consequences of your words or actions.
- Stay aligned with your values under pressure: Do not let platform incentives, audience reactions, or algorithmic rewards dictate how you show up.
QUESTIONS TO ASK BEFORE ACTING
- Am I responding to be helpful, or to be seen?
- What is the possible impact of my action, even if my intention is positive?
- Does this contribution open space for others, or close it down?
- If someone names harm, am I prepared to receive that without defensiveness?
- What does accountability look like for ME in this situation?
- What part of my response is grounded in my values, and what part is shaped by the platform’s algorithm?
EDEN is the starting point, not the finish line.
This guide introduces the language, principles, and structure needed to move through AI-shaped spaces with clarity, responsibility, and agency. It sets the foundation, but the full vision reaches much farther.
WHAT COMES NEXT FOR EDEN
The ideas introduced here are the foundation for a much broader body of work. The next phase expands EDEN from a framework into a living practice:
- Shared practices that help people apply these norms in real conversations, real products, and real organizational decisions.
- A community space where people learn together, compare experiences, hold accountability, and shape new norms that match the realities of AI-driven life.
- A library of tools and trainings that translate EDEN’s principles into practical steps for leaders, creators, teams, and institutions.
- A learning ecosystem that evolves through collaboration, research, experimentation, and real-world application across industries.
For now, this guide gives you the core: a grounded understanding of power, identity, autonomy, and the responsibility each of us carries inside AI-driven environments.
The next phase will turn these ideas into action, whether through workshops, communities of practice, or deeper guidance for organizations trying to lead with integrity.
This is only the beginning.
Last Update: 12/2/2025
If you’re reading this and thinking your team or organization needs support navigating AI with more intention, reach out. I’m already advising leaders who want to engage ethically, strengthen alignment, and make responsible choices in this landscape.
If that’s the direction you’re moving, I’m open to a conversation about what support could look like. Feel free to send an email or schedule a quick call.