As AI reshapes how we create, work, and share knowledge, what standards of care will protect our stories, intellectual property, and humanity?
Artificial intelligence is transforming how we create, communicate, learn, and work. The opportunities are extraordinary, and the pace of development is staggering. As quickly as we become familiar with one tool, another promises to transform our lives through autonomous agents, humanoid robots, driverless vehicles, and increasingly sophisticated systems. With so much activity devoted to feeding, training, and improving AI models, often for good reason, I keep returning to a question: How do we harness AI to do our best work without allowing it to colonize our stories, knowledge, and intellectual property? I recently experienced something that made this question deeply personal.
When AI Gets Our Stories Wrong
While using an AI application to develop an animated video about an experience from my high school years, I requested images of a 16-year-old Black girl, first at school and then participating in a mock trial courtroom. Despite repeatedly specifying her age, setting, and appearance, the application continued generating adultified, sexualized versions of a Black teenage girl wearing tight, short, and revealing clothing. I tried the same prompts with girls of other races. The results were strikingly different.
The other teenage girls appeared in age-appropriate clothing, including modest suits for the mock trial setting. The Black girls were repeatedly depicted in revealing attire, regardless of my instructions. This was a disturbing example of the racialized and sexualized stereotypes that Black girls have endured for generations, now reproduced through a technology capable of generating and distributing those representations at scale. I contacted the application's creators. My concerns did not lead to meaningful progress that I could see.
The experience left me thinking about the information used to train these systems, whose perspectives influence their development, and who has the authority to challenge the representations they produce. Black girls deserve to be represented with the same dignity, innocence, complexity, and age-appropriate care afforded to other children. That should be a fundamental expectation of any technology generating images of children. I wanted the model to draw from stronger, more culturally accurate sources. I wanted the people developing it to recognize the harm and correct it. Most importantly, I wanted to know what standards were in place to prevent it from happening again.
The Questions I Carried Into My Work
I carried that experience with me as I began working with clients in curriculum development, special education, and educational technology. As a former educator who has coached teachers and partnered with school leaders on talent operations, I understand the responsibility that accompanies decisions affecting students, families, and learning environments. My current work raises new questions about how we exercise that responsibility when AI becomes part of the process.
Even with expertise in our respective fields, my colleagues and I have had to ask ourselves: What standards of care and contextual integrity are we applying to the content we develop? How do we evaluate accuracy, cultural representation, and the potential consequences of the information we produce? What protections should guide our use of student and client data? These are questions I continue to ask as a creative professional and business owner. My standards of practice are evolving as I learn more about the technology, its possibilities, and its limitations.
I want AI to help level the playing field for people and communities historically underrepresented and underserved. I want it to make our work more efficient, free our time and energy for higher order thinking, and help us move ideas forward so more people can access the support they need, sooner and with greater accuracy.
Is that happening? In some ways, yes. Yet alongside these advances, Black girls are being grossly mischaracterized, people are losing money to AI enabled impersonation and fraud, musicians and artists are confronting unauthorized imitation and reuse of their work, and manipulated stories are circulating as mixtures of truth and fabrication. Our privacy, intellectual property, and capacity for independent judgment are also being tested. The consequences are far too significant to leave standards of care as an afterthought.
Who Gets to Tell Our Stories, and Who Gets to Profit?
Our stories are how we articulate our world, preserve knowledge, name our experiences, guide relationships, and understand who we are. Long before AI, communities were confronting historical and contemporary systems that devalued, appropriated, and exploited their knowledge, cultural practices, and lived experiences. AI introduces another dimension to these struggles. Information can now be collected, analyzed, reproduced, and commercialized at extraordinary speed, often without the people who created it fully understanding how it will be used or who will benefit.
I recently learned of a programmer whose contributions helped improve an AI model that reportedly generated millions for its owners, while she received a comparatively small payment and later found herself at odds with how her work was being used. Some might respond that this is simply how business works. Did she read her contract? Did she negotiate her compensation? Did she establish boundaries or ask enough questions?
I would offer a question in return: Have we?
How many of us understand the agreements we accept when uploading our work, connecting our accounts, or contributing information to systems we do not control? How much influence do we retain over what happens next? These concerns extend beyond contracts and compensation. They reach into questions of cultural ownership, community authority, consent, and the responsibilities that accompany working with other people's knowledge.
Indigenous storywork scholars, including Jo-ann Archibald Q'um Q'um Xiiem, Jenny Bol Jun Lee-Morgan, and Jason De Santolo, offer important perspectives on respect, reciprocity, cultural knowledge, and the relationships between storytellers and their communities. Their work provides a valuable foundation for examining the responsibilities that accompany the collection, interpretation, and use of stories.
Within AI research, Timnit Gebru examines structural harms and power imbalances in AI development, Joy Buolamwini has exposed racial and gender disparities in facial analysis technologies, and Rumman Chowdhury advances approaches to AI auditing and testing for harmful outcomes. These scholars and practitioners, alongside many others, are helping us examine the systems, practices, and protections needed to make AI more accountable.
When Innovation Moves Faster Than Accountability
Driverless cars, humanoid robots, and autonomous agents promise to transform how we live and work. Each advancement brings new possibilities, yet the speed of development is challenging our ability to establish, understand, and enforce adequate safeguards. Even as certain tasks become easier for many of us, we remain collectively outmatched in technical capacity and access to information, except for a relatively small group with specialized expertise and influence.
Doesn't that sound familiar? A familiar concentration of power, reimagined for the age of AI. We should be asking who gets to shape these technologies, whose interests are prioritized, and who bears the consequences when things go wrong.
The National Institute of Standards and Technology has developed an AI Risk Management Framework. UNICEF, UNESCO, the U.S. Copyright Office, and other institutions are examining protections involving children, education, intellectual property, and responsible AI development. This work matters. So does our ability to distinguish meaningful safeguards from gestures that merely create the appearance of accountability.
I am less interested in cosmetic assurances of transparency, such as a watermark identifying AI generated content, when the larger questions of exploitation, security, consent, and accountability remain unresolved. A watermark may help identify synthetic content. It cannot prevent someone's life savings from being stolen through impersonation, protect a family from benefits fraud, or guarantee that a community's cultural knowledge will be treated with respect. We need standards of care that address the actual risks people face, with protections proportionate to the potential harm.
Before We Opt In, What Standards of Care Should We Expect?
What standards should guide our use of AI when it comes to personal and family information, student and client records, business intelligence, creative work, and intellectual property? What about the wisdom of our ancestors, our cultural heritage, and our innermost thoughts? The questions vary across sectors. Parents must consider how AI companions interact with adolescents. Educators need to protect student information and evaluate how AI influences learning. Entrepreneurs and artists must understand their rights to original work, while healthcare providers, employers, and public institutions must account for the consequences of inaccurate or discriminatory decisions.
Some may choose to avoid certain AI products altogether, although complete avoidance is increasingly difficult as these systems become embedded in everyday services. For those of us participating, knowingly or unknowingly, we need practical ways to assess the technologies we use and the protections we should expect.
Inspired by established principles of responsible AI, including the NIST AI Risk Management Framework, I've organized 7 questions to consider before we click “Agree”:
Purpose: What am I asking this technology to do, and is it appropriate for the task?
Permission: Do I have the right to share the information, images, stories, or intellectual property involved?
Privacy: What information is collected, retained, shared, or used to train these systems?
Protection: What safeguards exist against misuse, exploitation, discrimination, and unauthorized access?
Proof: How can I verify the accuracy, authenticity, and cultural integrity of what the system produces?
Power: Who benefits, who carries the risks, and who is accountable when something goes wrong?
Preservation: What control do I retain over my information, creative work, and contributions?
These questions are a starting point, not a substitute for the responsibility of developers, companies, and institutions to establish and enforce meaningful safeguards. They can, however, help us make more informed decisions, advocate for stronger protections, and exercise greater care with the information, knowledge, and stories entrusted to us.
Keep Asking. Keep Learning. Help Shape the Standards.
As we celebrate each new tool, agent, and technological breakthrough, I want us to be equally invested in understanding how we can coexist with these technologies while protecting what matters most: our creativity, intellectual property, privacy, cultural knowledge, relationships, and ability to define ourselves. As I continue learning and refining my own practices, I've discovered researchers, scholars, and practitioners doing important work to establish standards, investigate harms, and strengthen accountability. Their work offers valuable guidance as we develop questions and practices informed by our own professions, experiences, and communities.
To support that exploration, I've assembled 2 companion resources:
People to Study: Exploring AI Standards of Care highlights scholars, researchers, practitioners, and organizations advancing work in AI accountability, intellectual property, privacy, Indigenous storywork, cultural representation, and responsible technology.
Before You Opt In: Questions to Ask About AI provides practical questions for families, educators, entrepreneurs, creatives, healthcare professionals, technologists, government agencies, and community organizations.
Neither resource is comprehensive. I encourage you to explore additional perspectives, develop questions relevant to your own circumstances, and share what you learn. The more informed and engaged we become, the better positioned we are to help shape the standards that protect us all.
