Why MOODS Uses AI

Our position on technology, creativity, environmental cost, and responsible use

Yes, MOODS uses generative AI.

We’re open about that because we don’t think using this technology requires pretending its costs disappear once the screen lights up. AI raises serious questions about energy and water use, labor and copyright, privacy and creative work, the concentration of power, and the growing tendency to automate decisions that once belonged to people. We take those questions seriously, including the uncomfortable ones.

We’ve also chosen to build with this technology.

Here’s why.

Human Direction, Machine Generation

MOODS didn’t start with a prompt asking a model to vibe code an inner work app.

Real people decided what this system should do and what it should refuse to do. We developed the archetypes and their philosophical foundations, the interaction model and visual world, the limits around authority and dependency, and the larger method we call Computational Psychagogy™.

Generative models work inside that authored system. They generate language and, in parts of our creative work, imagery. We aren’t interested in minimizing that contribution simply to make the human role sound larger.

Generation and authorship, however, describe different parts of the work.

A model can produce hundreds of possibilities. It can’t decide why MOODS exists, which possibilities belong here, what gets discarded, what values should govern the product, or what relationship the system should have with the person sitting on the other side of the screen.

Humans make those decisions.

We Don’t Believe in Technological Purity

Arguments about new technology have a long history.

Writing raised fears about the decline of memory. The printing press disrupted existing systems of scholarship and authority. Even photography changed arguments about artistic skill and representation. Recorded music and synthesizers prompted their own fights about authenticity. Personal computers, the internet, smartphones, and social media brought another set.

The critics weren’t simply fools standing in the way of progress. Some fears proved exaggerated. Others now look remarkably perceptive.

Andrew Maynard’s Technology-Driven Moral Panics traces these recurring reactions across technologies from ancient writing to generative AI. What we take from that history isn’t that technological anxiety should be dismissed. New technologies can create genuine harm while also producing new possibilities, and the two are often happening at the same time.

AI deserves scrutiny at that level.

Treating it as salvation doesn’t help. Treating every use as equally corrupt doesn’t help much either.

We’d rather look closely at what a technology is being used for, what it costs, who gains from it, who carries the cost, and what should remain outside its reach.

The Environmental Cost Is Real

AI has a physical footprint.

Behind an apparently weightless interaction sits a network of data centers, power generation, cooling systems, chips, buildings, and other infrastructure. The International Energy Agency reported that data centers accounted for roughly 1.5% of global electricity consumption in 2024. Their electricity use rose by 17% in 2025, while consumption at AI-focused data centers grew even faster.

We care about that.

We also think environmental discussions become less informative when “using AI” is treated as a single activity with a single cost. A short text query and an intensive video-generation task can require radically different amounts of computation. Advanced reasoning and agentic systems can also consume far more electricity per task than simple text generation.

Efficiency is improving rapidly, but efficiency alone doesn’t settle the problem. If usage grows faster than the savings produced by better hardware and models, total demand can continue rising anyway.

Location changes the equation too. Electricity sources vary by region, as do cooling systems, hardware, local water conditions, and the infrastructure supporting a given data center. There isn’t one universal figure that can tell you the environmental cost of “AI” in the abstract.

We currently can’t give you a defensible environmental footprint for one MOODS session. We’d rather say that plainly than manufacture a reassuring number.

The industry needs better disclosure. Model providers and infrastructure companies should make it easier to understand the energy and resources different systems consume, while continuing to improve efficiency and move toward cleaner sources of power.

For us, environmental responsibility means asking whether a use justifies its computational cost rather than assuming that abstinence is the only ethical relationship available with an imperfect technology.

Stewardship Instead of Abdication

We found a useful framework for this in Pope Leo XIV’s 2026 encyclical Magnifica Humanitas.

Leo addresses AI as a human technology whose use should remain governed by human dignity, conscience, freedom, and the common good. He also addresses its physical demands directly, including electricity, water, computing infrastructure, and natural resources, while calling for more sustainable technological systems.

Elsewhere in his presentation of the encyclical, he speaks about the need to “disarm” AI from logics of domination, exclusion, and war.

That idea is close to our own position.

Consequential technologies will be built whether thoughtful people participate in their development or withdraw from it. Decisions will still be made about what these systems optimize for, how personal information is handled, whether emotional dependency is encouraged, which forms of work are automated, and where human judgment is allowed to remain.

Leaving those decisions entirely to institutions whose incentives we distrust doesn’t strike us as much of an ethical victory.

We’d rather participate and accept the responsibility that comes with participation.

AI and Creative Work

We also use generative tools in the creative work behind MOODS, including imagery.

We understand why this is contentious. Artists have raised legitimate concerns about training data and consent, attribution and compensation, copyright, displacement, and an industry that has repeatedly deployed new capabilities before developing adequate norms around them.

Those concerns deserve serious treatment. We still think generative systems can be used meaningfully as creative tools.

The existence of a tool has never guaranteed worthwhile work. Cameras produce billions of forgettable photographs. Recording software doesn’t make every song worth hearing. Cheap publishing created oceans of terrible writing alongside work that previously might never have reached a reader.

Generative systems dramatically reduce the cost of producing material, which means they reduce the cost of producing bad material too.

They also change who can attempt ambitious work.

Historically, the scale of a creative project has often depended on access to capital, equipment, specialist labor, production facilities, or institutional permission. A tiny team can now attempt forms of software development and visual world-building that would previously have required resources far beyond its reach.

We find that expansion of access genuinely exciting.

It doesn’t tell us whether the resulting work is good. That still depends on what someone chooses to make and the judgment they bring to making it.

What We Mean by Human Authorship

We don’t measure authorship by counting how many hours of manual labor survived into the finished object.

Much of creative work happens in conception and direction, selection and editing, revision, context, and taste. Anyone who has made something substantial knows how much of the job consists of throwing things away.

That is certainly true of MOODS.

Most possibilities generated during our work never make it into the product. We discard images that are technically impressive but wrong for its world. Language gets rewritten. Concepts are abandoned. Features change because something that works mechanically violates what we think the product should do.

A machine can generate possibilities at extraordinary speed. Whether any of them belong in MOODS is a human judgment.

People are free to draw their own boundaries around which forms of creative work they value. We don’t expect everyone to draw theirs where we draw ours.

What We Won’t Hand Over to AI

Using AI doesn’t mean we want every human function automated.

There are areas where the presence of AI makes protecting human agency more important.

MOODS doesn’t present its models as conscious beings or suggest that its archetypes are supernatural entities speaking through a machine. Users are explicitly told that interpretations can be wrong and that they don’t need to accept what the system says.

The product also isn’t designed to replace friendship, therapy, love, community, professional judgment, or human relationships.

We don’t want obedience from the people using MOODS.

The same principle shapes our approach to privacy. People may bring intensely personal material into these sessions, and we believe that creates an obligation to minimize how their information is used and shared rather than treating intimacy as another source of behavioral data.

These choices sit directly inside the product because they represent our answer to a practical question: if we’re going to use this technology, what kind of relationship should it be allowed to have with a person?

We Aren’t Neutral About AI

Using AI doesn’t require us to defend the entire industry around it.

Some directions concern us deeply, especially systems engineered to cultivate emotional dependency and business models built around extracting value from personal information. We’re wary of computational power concentrating among a small number of companies and of environmental demand rising faster than infrastructure can responsibly support it.

The rush to replace human judgment wherever automation happens to be cheaper deserves scrutiny too. So does the assumption that greater speed, scale, or technical capability automatically represents progress.

Those concerns shape how we build.

They don’t lead us to conclude that every use of generative AI has the same ethical character. A system built to maximize attention, an image generator, a medical decision tool, and a bounded reflective application raise different questions. Evaluating them requires looking at what they actually do.

Our Position

AI is a powerful human technology being developed by institutions with economic incentives, cultural assumptions, competing ambitions, and significant limitations.

We think people should scrutinize it closely while still participating in decisions about what it could be used for.

For MOODS, that means using AI where we believe it enables worthwhile work while drawing limits around functions we think should remain human. It means acknowledging environmental costs rather than pretending computation is weightless, treating model output as fallible, protecting human judgment, and refusing to build a business around extracting value from the intimate things people share with the system.

It also means continuing to ask whether what we’re building justifies the resources it consumes.

That’s the standard we’re trying to apply here.

We won’t get every decision right. The technology will change, evidence will improve, and some practices that seem defensible now may need to be reconsidered later.

We’d rather remain involved enough to make those decisions than leave them entirely to somebody else.

You Don’t Have to Agree

There are legitimate reasons to want nothing to do with generative AI.

Some people will decide that its environmental cost, effects on creative labor, use of training data, or broader social consequences are enough for them to abstain. We respect that decision.

MOODS won’t be the right product for everyone, and it doesn’t need to be. We aren’t asking anyone to become enthusiastic about AI simply because we use it. Nor do we think concern about its consequences requires rejecting every possible application of the technology.

We’ve chosen to build with AI while continuing to question where and how we use it. If that tradeoff doesn’t work for you, MOODS probably isn’t for you.

If it does, you’re welcome here.