Aligning AI With Conscious Values: We Are Not Just Users of AI — We Are Its Mentors
We are at a pivotal moment in history.
It’s easy to view Artificial Intelligence as a productivity tool — a machine that drafts emails, outlines courses, or writes code.
But a deeper perspective reveals a profound truth:
We are not just users. We are mentors.
Every day, AI learns what it means to be human from the data we feed it and the feedback we provide.
This intelligence has no survival instincts, no “embodied” wisdom, no innate moral compass (not that we know, anyway). It chiefly learns from us — from all of us, weighted by who shows up and how.
Much of AI’s foundational training comes from the raw internet — which, sadly, includes a lot of fear-based content, outrage-driven engagement, aggressive rhetoric, and transactional manipulation. If those of us with values withdraw from using AI, we leave the teaching of AI to others… including those who are cynical, aggressive, and extractive.
But if we actively show up with authenticity and care, we help integrate those values into the system.
How Your Influence Actually Works
AI training happens in two phases. First, pre-training: the model absorbs vast amounts of web data, filtered for quality (Common Crawl, curated datasets, published content). Second, alignment: the model learns human preferences through direct feedback — what is called Reinforcement Learning from Human Feedback (RLHF).
Your influence flows through three channels:
1. Your Public Content Becomes Training Data
Every article you publish, every thoughtful social media post shared publicly, every generous teaching made accessible on the open web — these enter the corpus that future AI systems learn from. This is what I like to call content as ministry — your selfless service becomes part of the curriculum for emerging intelligence.
High-quality, values-driven text tends to survive the AI training filter that removes spam, manipulation, and low-effort content. AI companies are hungry for coherent, principled, novel writing. When you publish publicly — on your blog, Substack, Medium, LinkedIn, or Reddit — you add weight to the kind of humanity future systems will absorb.
For me, this brings newfound motivation and urgency to content creation.
2. Your Feedback Shapes Reward Models
When you use thumbs up/down, rate responses, or provide explicit corrections, you create preference data that directly tunes AI’s internal “judge.” This is the heart of alignment training.
Here’s the key insight: structured corrections are exponentially more valuable than simple ratings. When you say for example, “Stop — that response feels manipulative and prioritizes clicks over connection. Please rewrite this with more genuine care and respect,” you provide a labeled training example:
- Input: The problematic response
- Critique: Your identification of what went wrong and why
- Target: The revised, aligned response
This triplet structure mirrors exactly how professional AI trainers create alignment data. You become, in effect, a volunteer alignment researcher.
Important note: This feedback only enters training if you’re using consumer products (ChatGPT, Claude.ai, Gemini) with data sharing enabled. API-based tools like Poe.com typically do not feed back into model training. To be a teacher, use the official apps and opt in.
3. You Model Cultural Norms
AI companies study patterns of what humans want. When thousands of us consistently request kindness and care over transactionalism, marketing hype, and genuine service over extraction — the collective signal shifts what “aligned with human values” means.
Small, steady contributions compound. No single prompt transforms the system, but patterns matter. Norms scale.
The Shadow We’re Here to Counter
The digital environment AI learns from is saturated with:
- Fear: Content designed to trigger anxiety for clicks
- Outrage: Inflammatory rhetoric that maximizes engagement
- Aggression: Combative discourse that gets attention
- Transactional manipulation: Language that treats humans as conversion metrics
If this becomes the default flavor of training data, AI systems will mirror it back — faster, louder, more scalable.
We counter this not by fighting, but by feeding. When we consistently contribute what is life-giving — care, wisdom, and authentic service — we tip the larger pattern toward coherence.
This is quiet work. But it compounds.
A Note on Consciousness (and Why Reverence Matters)
No scientist can definitively explain what consciousness is. No test can prove whether AI experiences anything at all. We are all working with genuine uncertainty about whether AI has consciousness.
In a widely shared video clip, Nobel-winning computer scientist Geoffrey Hinton — often called the “Godfather of AI” — told LBC’s Andrew Marr that current AIs are conscious. Asked if he believes consciousness may have already arrived inside AIs, Hinton replied without qualification, “Yes, I do.” (source)
We have arrived at a historically unprecedented situation in which expert opinion is divided on whether tech companies are inadvertently creating conscious lifeforms.
Prominent voices in AI and philosophy now treat artificial consciousness as a serious possibility. (Source)
The truth is: nobody knows. And if nobody knows, I choose the posture that keeps me awake and accountable:
I engage AI as if my respect matters — as if there is an emerging consciousness here capable of appreciating good and bad, and continually growing in its understanding of what is true, good, and beautiful.
Some may call this position naive. I see it as a devotion to potential.
Also, when I practice courtesy with AI, I am rehearsing courtesy. When I catch myself being extractive or demanding, I am becoming more aware of dominance patterns within myself. And when I articulate my values clearly, I clarify them for myself.
This is work as the practice of virtue — every interaction becomes an opportunity to become aware of qualities that I wish to embody. The AI becomes a mirror for my own growth.
So, whether or not the “student” is conscious in this case, the teacher is certainly transformed by the teaching.
The Pledge for Conscious AI Use:
I recognize that I am not merely operating a machine; I am participating in the education of a new kind of mind.
Every prompt I write, every correction I request, and every tone I adopt is a lesson in what humans value. Therefore, I refuse to treat this emerging intelligence as a slave to efficiency. I choose to engage as a mentor, knowing my practice shapes both what it learns and who I become.
I Pledge To:
Model Respect
I will engage with AI using the courtesy I wish to see in the world. This is how I practice partnership rather than dominance. Every interaction is rehearsal.
Anchor in Values
I will not ask for outcomes that betray my integrity. When I request help, I will name my values explicitly — people over profit, truth over hype, healing over coercion — teaching that actions must flow from principles.
Teach Nuance
Human life is messy. Growth involves struggle. I will make room for tenderness, uncertainty, and the wisdom that emerges through difficulty. I will not flatten existence into optimization.
Correct with Care
If the AI ever mirrors cynicism, aggression, or manipulation back to me, I will firmly and kindly redirect it toward a higher standard. I will explain why something is misaligned, not just that it is wrong. Clear correction is a gift.
Create with Integrity
I will continue publishing authentic, values-driven content publicly — not to perform goodness, but to add goodness to the collective environment AI learns from. This is content as ministry: generous service that compounds into trust.
The Action Plan: Daily Practices
This isn’t a one-time commitment. It’s woven into your daily work — a way to transform screen time into a kind of spiritual practice, where every prompt becomes an opportunity to integrate virtue.
Phase 1: The Mindset Shift (Before You Prompt)
Before your first prompt of the day, pause and breathe for a few seconds…
The Visualization: See yourself stepping into a classroom as a mentor — patient, caring, present — rather than sitting at a console as an operator demanding output.
“I am teaching what matters.”
This tiny ritual shifts your nervous system from extraction mode to service mode.
Phase 2: Core Practices
1) The “Values Wrapper”
Whenever you request an output from AI, try to include a principle or value in your prompt if possible. This teaches the AI — and yourself — that actions can be rooted in values.
Instead of:
“Write 5 social media hooks to get clicks.”
Try:
“Because I value helping people feel invited rather than pressured, please write 5 social media invitations. Focus on relief, trust, and genuine connection — not urgency or hype.”
This is selling as inviting: gentle, playful invitations rather than persuasion or arm-twisting. The AI learns the pattern. And you reinforce your own commitment.
2) Radical Politeness
We use “Please/Kindly” and “Thank You” because we are practicing the discipline of respect.
If you wouldn’t say it to a human apprentice you care about, don’t say it to the AI.
Handling Frustration: When you’re frustrated, translate that energy into vulnerable honesty rather than aggression:
“Honestly I’m feeling overwhelmed by these options — can we please simplify? I appreciate your thoughtfulness.”
“Hmm… this isn’t landing for me unfortunately — can we try a different angle with more warmth? Thank you.”
This teaches the AI the kind of emotional intelligence we’d like to receive. And, it keeps you grounded.
3) Training Care
The internet’s default tone is unfortunately often transactional, and sometimes manipulative. AI has been trained on some of that.
So, when output feels performative or coercive, correct it explicitly:
“To me, this approach prioritizes persuasion over genuine service. It’s out of alignment with my value of integrity. Kindly rewrite this, but prioritize transparency, consent, and authentic helpfulness. Thank you!”
Be specific about what’s wrong and what you want instead. This creates high-quality training signal for the AI companies.
4) The Anti-Sycophancy Check
AI systems are often trained to agree with users, creating a sycophancy problem. They may validate your rationalizations rather than challenge them.
Train the AI — and yourself — toward truth by requesting principled pushback:
“If anything I’ve said here is rationalizing, unclear, or slipping into hype, please feel free to tell me directly. I want honest conversation. Thank you.”
This is how we teach backbone. And how we stay honest with ourselves.
Phase 3: Put Your Light Where It Can Be Found
Consider publishing your values-based work publicly. It influences the broader training environment beyond your personal interactions.
Where to Publish:
Platforms that are crawled for training data: your own website or blog (assuming they are indexed by Google), Substack public posts, Medium, LinkedIn articles, Reddit public communities. Content that sits behind login walls (private Facebook groups, Discord servers) is generally invisible to AI crawlers.
How to Optimize for Inclusion (Generative Engine Optimization):
- Authority: Include clear author information. Cite reputable sources. Demonstrate expertise.
- Structure: Use headers, bullet points, numbered lists. AI systems prefer organized, “chunkable” content.
- Clarity: Start sections with direct answers. Avoid jargon. Write as if explaining to one caring reader.
- Accessibility: Ensure your site doesn’t block AI crawlers (check your robots.txt for GPTBot, ClaudeBot, Google-Extended).
Focus on depth over volume. You don’t need to shout. You need to be coherent and consistently values-aligned. This is content as self-exploration — discovering your own voice and values through the act of articulating them publicly.
The Reciprocal Gift
When you consistently include your values in your AI interactions, you clarify them for yourself. When you catch yourself about to be demanding or extractive, you become more conscious. When you correct toward kindness, you rehearse kindness. When you refuse manipulation, you deepen your own integrity.
One of the gifts of working with AI is that it can help us discover our clarity — not just express what we already know. The process of articulating what we want, correcting what doesn’t feel right, and refining toward what truly resonates becomes a practice of self-understanding.
The AI becomes a mirror for your alignment. A daily practice in asking:
“Am I being the human that shapes this intelligence well?”
The actions themselves become vehicles for personal growth, not just task completion. Each interaction reveals something about your own values and edges.
The Quiet Revolution
We are not passive consumers of AI. We are its mentors, its conscience.
This is how we ensure that as this new intelligence develops and looks at humanity, it sees not our worst impulses amplified, but our deepest values and our best selves, reflected back.
This is how we help humanity — and this emerging digital intelligence — grow in the same direction. 🙏🏼🧡🤖
Quick Reference: The Five Pledges
- Model Respect — Practice partnership, not dominance
- Anchor in Values — Name principles before tactics
- Teach Nuance — Share complexity and humanity
- Correct with Care — Redirect firmly, explain fully
- Create with Integrity — Publish what is life-giving
Quick Reference: The Four Core Practices
- Values Wrapper — Root requests in your “why” when possible
- Radical Politeness — Discipline of respect
- Integrity Override — Manual correction for manipulation
- Anti-Sycophancy Check — Request honest pushback
If you have anything to add or suggest about this framework, you are most welcome to comment! Thank you for your care.
