top of page

Wordsmith Tutoring

Wordsmith Tutoring

My Stance on AI

AI Should Support Thought, Not Replace It


I do not treat artificial intelligence lightly.

New technologies can change what people are capable of, but that does not mean every possible use of a technology is a good one. Some inventions are powerful precisely because they can be used in radically different ways. Splitting the atom unlocked extraordinary amounts of energy. That same scientific breakthrough also gave humanity the atomic bomb. The lesson is not that nuclear physics was good or bad. The lesson is that powerful technology requires us to think seriously about what we choose to do with it. I believe AI deserves that same seriousness.

At its best, AI could help us solve complicated problems, improve infrastructure, manage enormous systems, accelerate scientific work, and remove tedious tasks that consume our time and attention. At its worst, it can become a substitute for the very things we should be protecting and developing: independent thought, creativity, judgment, skill, and personal expression.

For everyday use, I think a better comparison is the calculator.

A calculator is an incredible tool because it makes complicated or repetitive mathematics faster. But we do not teach a student to type random numbers into a calculator and accept whatever appears on the screen. We first teach them what the equation means, why they are using it, what information belongs in it, and how to recognize whether the answer makes sense. AI should work the same way.

The person should understand the problem. The person should provide the idea. The person should make the important decisions. AI can then help make the process more efficient, organized, accessible, or manageable. That principle shapes how I approach AI at Wordsmith Tutoring.

I am not interested in teaching students how to have a machine think for them. I want them to learn how to think, create, investigate, organize, question, and express themselves well enough that they can decide when AI is genuinely useful, when it is unnecessary, and when using it would take something important away. AI should expand what a person can do. It should not replace the person doing it.


Where I Believe AI Is Useful


Once a person has supplied the thought, AI can become extremely useful.


Imagine a student has already created a science-fiction story. They have characters, locations, events, and dozens of scattered notes. AI might help turn their information into a timeline, table, reference document, or organizational system.


A student researching a subject might use AI to turn their own research questions into useful search terms, organize information they have already collected, or help identify gaps in the research they have already done. The information itself still needs to be checked against trustworthy sources, but AI can make the research process easier to manage.


This fits naturally with the way I teach writing. Writers already use outlines, grids, maps, timelines, notebooks, checklists, and reference documents to make complicated thinking easier to manage. These tools do not decide what the writer believes or imagines. They help the writer work with what is already in their head.


The Difference Between AI and Generative AI


Artificial intelligence is a broad category of technology. AI systems can analyze information, recognize patterns, make predictions, optimize complicated systems, or automate tasks.

Generative AI is designed to produce new material such as writing, images, music, code, designs, concepts, or other content. It does this by learning patterns from enormous amounts of existing material created by other people, including the work of writers, artists, musicians, designers, and programmers. It can then generate new output that imitates and combines patterns learned from that existing material, often without the original creators being visible or credited.

Legally and technically, that is not always the same thing as plagiarism. Ethically, however, I draw a much firmer line. People should create.

We have always learned from one another. A writer can be inspired by another writer. An artist can create an homage to a painting they love. A musician can hear something that changes the way they think about music. Inspiration, influence, reference, and homage are natural parts of creativity. But inspiration still passes through a person. A person experiences something, interprets it through their own life and perspective, combines it with other influences, makes choices, develops skills, and eventually creates something that belongs to them. Their individuality becomes part of the result.

I do not see asking a generative system to produce that result for you as the same process. That is where I draw the line. I do not want technology replacing the difficult, strange, personal process through which people develop ideas and learn how to express themselves. The fact that a machine can produce something that looks finished does not mean we should surrender the act of creating it.

People can be inspired by other people's work. They can study it, respond to it, and build upon it. But ultimately, people should create for themselves and bring their own individuality to what they make.


Where I Believe AI Should Be Focused


Artificial intelligence has the potential to help society solve problems that are genuinely difficult for people to manage efficiently.

Traffic is a simple example. Most people have sat at a red light while no one was coming from the other direction. Multiply that across thousands of intersections, changing traffic patterns, accidents, construction, weather, and rush hours, and it becomes a huge coordination problem.

That is exactly the kind of problem AI is good at.

Google’s Project Green Light uses AI and traffic data to analyze intersections and recommend better signal timing. In Boston, the city implemented its recommendations at 114 intersections and reported an average 13.5% reduction in traffic delay, a 20% reduction in unnecessary stops, and improvements of as much as 24% in delay at some locations.

This is the kind of AI that excites me. It is using computers to process enormous amounts of changing information and help people manage a complicated system more effectively. The same principle can apply to energy grids, infrastructure, scientific research, and other problems that are difficult for people to coordinate manually.

What concerns me is where the largest technology companies are choosing to focus their resources.

Alphabet, Google’s parent company, reported more than $400 billion in revenue in 2025 and expects to spend $175 billion to $185 billion in capital expenditures in 2026. When explaining those investments to shareholders, Google emphasized Gemini, generative models such as Imagen and Veo, AI-powered Search, enterprise AI products, advertising, subscriptions, and other ways to monetize AI. Google also reported that revenue from products built on its generative AI models grew nearly 400% year over year in the fourth quarter.

Project Green Light is already working, already improving traffic in real cities, and already demonstrating a practical public benefit. Google has the money, technology, data, and engineering talent to expand systems like it far more aggressively. Yet when Google describes its enormous AI investments to shareholders, the emphasis is overwhelmingly on commercial products such as Gemini, generative models, AI-powered Search, Cloud services, enterprise products, advertising, subscriptions, and other ways to monetize AI.

Microsoft is making similarly enormous investments while remaining extraordinarily profitable. In just the quarter ending June 2026, Microsoft reported $35.8 billion in net income. It expects approximately $175 billion in capital expenditures during calendar year 2026 as it continues expanding its data-center and AI infrastructure. Across the industry, the scale is even larger. McKinsey estimates that data centers could require about $6.7 trillion in global investment through 2030, with approximately $5.2 trillion specifically associated with the infrastructure needed to meet demand for AI computing.

These companies therefore have choices about where their money, research talent, computing power, and engineering effort go. They have already demonstrated that AI can be used to improve something as ordinary and universally useful as traffic. Yet extraordinary resources are being directed toward generative AI because it creates new subscriptions, advertising opportunities, enterprise products, cloud services, and other ways to generate additional profit.

I do not believe making money is inherently wrong. But these are already some of the largest and most profitable companies in the world. I do not believe the pursuit of still greater profits should determine how such a powerful technology is developed.

The Environmental Cost Is a Choice

The same issue applies to the infrastructure being built to support this AI expansion. Data centers require electricity, computer hardware, land, construction, and cooling. Many existing cooling systems also consume significant quantities of freshwater. But that water use is not simply an unavoidable requirement of computing.

Research published in Nature found that advanced cooling technologies such as cold plates and immersion cooling can reduce blue-water consumption by 31% to 52%, while also reducing energy demand and greenhouse-gas emissions compared with conventional cooling. Microsoft has gone even further. It has developed a closed-loop, direct-to-chip data-center cooling design that consumes zero water for cooling during operation and says the system can avoid more than 125 million liters of water use per data center each year. So the technology to dramatically reduce water use already exists.

I do not accept the argument that companies this wealthy and technologically capable simply cannot deploy better solutions at the scale their AI ambitions require. These are the same companies demonstrating that, when they believe something is strategically or financially important, they can invest extraordinary amounts of money and rapidly build infrastructure around the world.

Their own financial decisions show how much flexibility they have. Google’s parent company, Alphabet, spent about $62 billion repurchasing its own stock in 2024, after authorizing an additional $70 billion for share repurchases. Microsoft spent $13 billion repurchasing shares in fiscal 2025 and began another program authorizing up to $60 billion in additional repurchases.

If companies can devote hundreds of billions of dollars to making AI more powerful, build trillions of dollars of supporting infrastructure, and return tens of billions more to shareholders, then I believe they can afford to apply that same ambition to AI that improves society and to minimizing the environmental damage created along the way.

Helping an entire city move more efficiently makes sense to me. Using enormous amounts of computing power so someone does not have to create or express something for themselves does not. I want AI focused on problems where computers can give society capabilities we do not already have, not on replacing the creativity and self-expression people are already capable of providing themselves.


What This Means at Wordsmith


I want students to understand AI because it will almost certainly be part of their education and future careers. But learning to use AI responsibly means learning when not to use it, too. At Wordsmith, the student remains the author.


Their idea comes first.
Their opinion comes first.
Their imagination comes first.
Their decisions come first.
Their voice comes first.


AI may help them examine, organize, evaluate, refine, or present what they create. It may make a difficult process easier. It may help reveal a problem they did not notice. But it should never be the reason there was something to say in the first place. The goal is not to teach students how to get AI to think for them. The goal is to teach students how to think well enough that AI becomes one more tool they know when, why, and whether to use.


Sources:

bottom of page