By Dr. Jasmin Cowin (Star Trek Fan)
Lieutenant Worf: “There are no options. The Prime Directive is not a matter of degrees. It is an absolute.” The Next Generation, “Pen Pals” (Season 2, Episode 15, 1989).
Early summer 2026 I decided to apply for the competitive Stanford University’s Fall 2026 Ethics, Technology, and Public Policy for Practitioners course. Every year, I take one synchronous course to be part of my transdisciplinary quest on knowledge and personal connection building. My final reflection and report on my course experiences are structured through the lens of Star Trek’s Prime Directive.
Captain Jean-Luc Picard: “We are all sworn not to interfere with other lives in the galaxy. If I save this boy, I break that law.” The Next Generation, “Justice” (Season 1, Episode 8, 1987).
Session one’s Opening Session and Moral Caffeination reading was Ursula K. Le Guin’s The Ones Who Walk Away from Omelas. I was unfamiliar with the short story of a prosperous society whose happiness rests on the deliberate and excruciating suffering of an imprisoned child. The reading troubled me, and I rejected the offered choices between accepting the child’s horrific suffering for the sake of communal happiness or walking away as cowardly solutions. Instead, I devised a third solution: an outsider who rescues the child. My inspiration: a personal hero of mine: Raoul Wallenberg, a Swedish rescuer in Budapest, who acted to save as many Jews as possible during a time when the Nazi regime normalized the persecution of its Jewish citizens.
Personally, I reject collective happiness as justification for a society’s deliberate disregard for individual suffering. A tyrant may experience genuine happiness while inflicting profound suffering on others. A society may celebrate its collective well-being while knowingly sacrificing the well-being of some of its members. But what does that happiness tell us about the ethical legitimacy of either? Nothing. Isn’t happiness a subjective experience? Aristotle’s eudaimonia and Kant’s conception of duty influence my ethical outlook. Society’s satisfaction cannot justify what it knowingly inflicts on someone powerless. Needless to say, I felt very righteous post-session and Dr. Beverly Crusher’s objection to Captain Picard supported my righteousness.: “It’s hard to be philosophical when faced with suffering.”
Dr. Beverly Crusher: “But aren’t we interfering either way? If we take no action, it’s a conscious decision to let the Boraalans die.” The Next Generation, “Homeward” (1994).
However, sessions 2: Algorithmic Decision-Making and Fairness; 3: Data Collection, Privacy, and AI; and 4; Political Economy of Technology and the Distribution of Power, especially the deeper exploration on automated hiring, gave Omelas a distinctive and contemporary institutional form. An algorithmic decision can be transparent and still produce unfair outcomes. I learned that disclosure does not confer accountability; neither does an audit remedy harm unless someone is empowered to act on its findings. My recurring question became: who possesses the authority and incentive to correct a system once its effects are visible? Our privacy session made this question more difficult to resolve – I was still looking for an Omelas closure. Lawrence Lessig’s account of law, markets, norms, and architecture helped me see technological defaults as a form of governance. I explored a related problem before the course in The Educational Data Ghost: AI, Learning Platforms, and European Sovereignty (July 2026). My article asked “whether students and faculty can know where their traces have gone” and argued for “the right not to be indefinitely reconstructed from one’s own traces.” Learning platforms retain evidence of student activity and faculty work long after an interaction ends. Failing to investigate those data practices is a decision with consequences, but intervention must respect consent, privacy, and legitimate educational purposes.
The lectures and fireplace presentations also made me reconsider the relationship between an AI system’s analytical moral competence and its behavioral moral acting. An LLM may identify a defensible action and explain Kantian duty without acting in accordance with either. When an agent has tools and permission to operate beyond a chat window, this gap becomes very dangerous as ethical fluency is not evidence of reliable behavior. I explored that topic in my October 1st Horasis’ article, Capabilities, Behaviors and Resistance to Oversight: How Much Autonomy Have We Already Given AI?, where I examined incidents involving autonomous agents and argued:
“Institutions need detailed activity records, clearly defined permissions, independent evaluations, and procedures for identifying and reporting unauthorized actions.”
Jasmin Cowin, Capabilities, Behaviors and Resistance to Oversight: How Much Autonomy Have We Already Given AI?
During our session on AI governance under uncertainty, Seth Lazar’s concept of anticipatory AI ethics challenged me to consider when emerging AI capabilities and their foreseeable harms create a duty to intervene before those harms materialize, even when the evidence remains incomplete and the legitimacy of preventive action must still be established. His technological horizon asks us to examine plausible capabilities and their effects within existing institutions, without pretending to predict society’s entire future. Consistent with the Star Trek framing I used earlier, I applied this approach in my article Where No Worm Has Gone Before: AI That Self-Replicates onto Your $55 TV Box , looking at AI that replicates across inexpensive consumer devices. I observed that while laboratories can test and constrain models, they cannot by themselves secure household hardware, regulate device manufacturers, or protect networks. The duty to respond must be distributed among actors with different powers. An earlier article of mine, What the OpenAI Cyber Defense Letter Says, Who Signed It, and What Remains Unclear (August 2026), posed a related problem. Examining a public warning hosted by a company with commercial interests in the proposed response, I wrote that “a warning can be self-serving and accurate in the same breath.” That remains important to my thinking. Recognizing a genuine cyber threat does not exempt the proposed intervention from scrutiny, particularly when access, benefits, and authority are concentrated among the organizations promoting it.
Captain Kathryn Janeway: “You have no idea what the consequences might be once you involve yourself.” Voyager, “Time and Again” (1995).
Sessions 5. AI, Child Safety, and the Burden of Responsibility and 6. AI & the Future of Work continued my journey to a more nuanced thought process. I revisited my ‘heroic’ and, in hindsight, thoughtless Wagnerian rescue of the Omelas’ child. My original position was that no community can claim moral authority to perpetuate suffering simply because its members collectively accept or benefit from it. Communal consent does not legitimize the injustice. However, after session 5, I started asking myself:
- Defensible aims: Ending the child’s suffering and simultaneously protecting Omelans’ rights.
- Proportionate means: Using methods appropriate to the harm being prevented, without inflicting unnecessary additional harm.
- Competence: Ensuring that those intervening possess the abilities and resources to do so effectively.
- Accountability: Accepting responsibility for the consequences, including the child’s and Omelas’ welfare after the rescue.
The difficult to bear child-safety session brought these distinctions into sharper focus. My writing on New Mexico v. Meta had questioned leaving child protection to platforms whose commercial incentives favor engagement. Yet proposed client-side scanning of private communications creates a different risk: a protective unleveraging of privacy rights to justify disproportionate surveillance.
The future-of-work readings in session 6 extended the question of inaction. The Stanford Canaries in the Coal Mine research suggests a loss of the entry-level tasks. When employers automate those tasks, who assumes responsibility for creating alternative routes into skilled work? Will GenZ become the “useless” new class with total economic irrelevance? (Harari, 2022)
Finally, as a professor preparing future teachers, I find myself considering a much longer horizon: how do we educate teachers today to prepare children for the world of 2060? We cannot reliably predict the professions, technologies, or social conditions those children will encounter. We can, however, prepare teachers to cultivate intellectual independence, critical inquiry, disciplinary knowledge, ethical judgment, and the ability to learn beyond existing frameworks. My concern is that increasing reliance on AI may deprive my current candidates of the productive, intellectual struggles and practical (and sometimes painful) experiences through which judgment and expertise develop.
Cumulatively, the Stanford course gave me a distinct philosophical vocabulary with my final take-away: Non-intervention is not neutral in education either. The choices we make in preparing teachers today will influence the intellectual autonomy and economic opportunities available to their students (and our children’s children) – a future yet to be lived.
