Build AI Around People.
We are entering a period in which artificial intelligence can fundamentally change how people work, learn, create, communicate, and solve problems.
Much of today's discussion focuses on productivity, efficiency, and reducing labor costs. Those benefits are real, but they represent only one possible direction for AI.
I believe we should ask a different question:
How can AI make people better at what they already do?
AI should remove friction, repetitive work, information overload, and unnecessary administrative burden so people can spend more time applying the abilities that make them human: judgment, creativity, empathy, leadership, curiosity, craftsmanship, and connection.
Augmentation Before Replacement
AI engineering should begin by looking for opportunities to augment human capability rather than immediately attempting to remove the human from a process.
This is not an argument against automation. Some work should be automated — dangerous, repetitive, highly mechanical, or unnecessarily administrative tasks are obvious opportunities. But automation should be a tool rather than the objective.
The objective should be improving human outcomes.
Old question
"How many people can AI replace?"
Better question
"How many people can AI empower?"
AI + Teacher, Not AI Instead of Teacher
Teachers carry substantial workloads beyond the time they spend directly teaching students. AI can assist with many of those tasks:
Without AI
With Human-Centered AI
The teacher continues providing the things AI cannot simply reduce to generated content: mentorship, judgment, encouragement, classroom leadership, understanding individual students, and human connection.
The Human Capability Multiplier
AI should multiply human capability rather than diminish human value.
Teacher
AI handles preparation and repetitive tasks.
Human focuses on students, instruction, and mentorship.
Doctor
AI helps organize information and identify patterns.
Human provides clinical judgment, accountability, and patient care.
Engineer
AI accelerates research, prototyping, documentation, and analysis.
Human determines architecture, constraints, tradeoffs, and direction.
Small Business Owner
AI helps with research, marketing, documentation, analytics, and routine operations.
Human focuses on customers, products, strategy, and relationships.
Creator
AI accelerates exploration and iteration.
Human provides intent, taste, experience, and meaning.
AI handles more of the mechanical burden.
Humans gain more capacity for meaningful work.
Principles for Human-Centered AI
Augment Before Replacing
Before designing a system to replace someone, determine whether AI can make that person dramatically more capable.
Automate Tasks, Not Human Worth
A job consists of many different tasks. Automating a task does not make the person performing that job obsolete.
Keep Humans Where Judgment Matters
AI can provide information, analysis, recommendations, and alternatives. Important decisions involving people should preserve meaningful human judgment and accountability.
Return Time to People
One of AI's greatest opportunities is giving people time back by reducing repetitive administrative and information-processing work.
Expand Access to Expertise
AI can make knowledge, education, technical assistance, and sophisticated tools available to people who previously could not easily access them.
Build Tools People Can Understand
People should understand what an AI system is doing, what information it uses, where its limitations are, and when human verification is appropriate.
Measure Human Outcomes
Success should not only be measured in tokens, model size, throughput, headcount reduction, or dollars saved. We should also ask: Did people become more capable? Did they gain time? Did they learn something? Did access improve?
Preserve Human Purpose
Technology should create opportunities for people to contribute, learn, create, and participate in society rather than designing systems around the assumption that human participation is an inefficiency.
AI is not the objective.
Better human outcomes are the objective.
LLMs, small language models, agents, robotics, computer vision, machine learning, and future AI technologies are engineering tools. The correct technology depends on the problem. We should not add AI to systems simply because AI is available.
The right approach
Not this
Efficiency Matters. People Matter Too.
Businesses should absolutely use AI to improve productivity, reduce waste, accelerate development, and remain competitive. But reducing headcount should not become the primary measure of AI success.
There is another economic possibility:
AI can allow the same person to accomplish more.
AI can allow small teams to compete with organizations that previously required enormous resources.
AI can give individuals access to capabilities previously available only to large corporations.
AI can help someone learn a new skill.
AI can help someone start a business.
AI can help someone build something they previously lacked the technical ability to create.
AI can make expertise dramatically more accessible.
That is not simply automation.
Human Leverage.
What I'm Building Toward
My work at joecairns.ai will explore AI from this perspective. I want to investigate how artificial intelligence, language models, small language models, agents, robotics, context engineering, and other emerging technologies can help individuals become more capable.
That means building projects, conducting experiments, documenting what works and what doesn't, and making that knowledge available to others.
The goal isn't to demonstrate AI for the sake of AI.
The goal is to answer a much more interesting question: What can a person accomplish when AI is designed to work with them?
The future of AI should not be humans versus machines.
It should be humans accomplishing more because of machines.
AI will undoubtedly automate work.
But automation alone would be an unimaginative use of one of the most powerful technologies we have ever created.
The larger opportunity is augmentation.
Giving people better tools.
Giving people access to knowledge.
Giving people new abilities.
Giving people time back.
And allowing individuals to accomplish things that previously required resources, expertise, or organizations far beyond their reach.
Build AI that makes people more capable.
That is the direction I believe is worth building toward.