Suki is a Japanese-American novelist whose debut explores the untranslatable spaces between languages, moments where English and Japanese diverge in ways that reveal cultural differences in how people think about time, obligation, and intimacy. The book receives critical acclaim in English.

Her publisher offers two options: commission five human translators for the most commercially viable languages over two years, or use an AI translation system to produce thirty language editions in three months. The AI translations are fluent and accurate. Early readers in test markets respond positively.

But when Suki reads the AI's Japanese translation, something is wrong. The passages she wrote specifically to exist in the gap between languages (where English grammar forces a directness that Japanese grammar allows you to avoid) have been smoothed over. The AI produced perfectly grammatical Japanese that misses the point entirely. The awkwardness was the art.

Suki's literary agent argues that reaching thirty markets outweighs the loss of nuance in any single translation. "Most readers won't know what they're missing." A translator friend disagrees: "What they're missing is the entire premise of your book. You wrote about the impossibility of perfect translation, and now you're distributing a product that pretends perfect translation is possible."

Suki compromises: she commissions human translators for Japanese, Spanish, and French, and uses AI for the remaining twenty-seven languages. But this raises its own question: are the readers in those twenty-seven languages receiving a lesser version of her work? And if so, is that acceptable in exchange for reaching them at all?

What do you think?

DISCUSSION QUESTIONS

• Is a wider but shallower reach more valuable than a narrower but deeper one?

• Does AI-mediated communication lose something essential, or is that loss acceptable for the sake of access?

• Should creators be transparent with audiences about which parts of their work were AI-mediated?

• Is there a meaningful difference between AI translating your work and AI generating your work?

• What happens to the professional communities (translators, editors, interpreters) that currently mediate creative exchange?

Nina runs a small design studio specializing in packaging for sustainable consumer brands. Her clients choose her specifically because she shares their environmental values. When she integrates AI tools into her workflow, the results are remarkable: she can explore ten times more concepts in the same timeframe, her iteration speed doubles, and her clients are thrilled with the quality.

Then a sustainability consultant she works with sends her a report. The cloud computing infrastructure her AI tools rely on consumes enormous amounts of energy and water. The consultant estimates that Nina's AI-assisted workflow has roughly tripled the carbon footprint of her design process. "You're making packaging for zero-waste companies using the most energy-intensive creative tools available," the consultant notes.

Nina investigates further. She learns that a single AI image generation session can consume as much energy as charging a smartphone dozens of times. Training the models she relies on required energy equivalent to the annual consumption of small towns. The data centers are cooled by millions of gallons of water, often in drought-prone regions.

But she also learns that her studio's total AI energy use is still a fraction of her clients' manufacturing energy. A colleague argues that obsessing over AI's energy consumption is "rearranging deck chairs" when transportation and manufacturing dwarf it. An environmental scientist counters that AI's energy use is growing exponentially and the time to establish sustainable norms is now, not after the infrastructure is locked in.

Nina considers switching to local, smaller AI models that use less energy but produce less sophisticated results. She wonders what she owes her clients, what she owes the environment, and whether these obligations are genuinely in conflict.

What do you think?

DISCUSSION QUESTIONS

• Should creative professionals factor the environmental cost of their tools into their practice — or is that an unreasonable burden?

• Is AI's environmental impact acceptable if the total is small relative to other industries?

• Does working on sustainability-focused projects create a special obligation to use sustainable tools?

• How should the industry weigh the environmental cost of AI against the time and resources it saves?

• Who should bear the cost of making AI infrastructure sustainable — tool providers, users, or governments?

Diego is a graphic designer working on a campaign for a community health initiative in a predominantly Latino neighborhood. He uses an AI design tool to generate imagery of families, healthcare workers, and community spaces. Every image the tool produces features light-skinned people in settings that look like affluent suburbs.

Diego manually adjusts his prompts, specifying skin tones, cultural markers, and architectural details specific to the neighborhood. The results improve but still feel off, the AI seems to have a narrower visual vocabulary for non-Western, non-affluent contexts. Poses are stiffer. Compositions are less natural. The tool is clearly better at generating some worlds than others.

Diego brings this up at a design conference. A machine learning researcher explains that the training data overwhelmingly represents commercially produced, Western-centric imagery because that is what dominates the internet. "The AI is a mirror," she says. "It shows us what we have already over-produced and under-produced."

A community organizer in the audience pushes back: "A mirror that only reflects certain people is not a mirror — it's a filter. And deploying that filter in healthcare communications has real consequences. People who don't see themselves in health materials are less likely to seek care."

Diego is forced to weigh his options. He can spend extra time fighting the tool's defaults. He can commission a local photographer instead and skip AI entirely. Or he can accept the imperfect AI output for efficiency and focus his effort elsewhere. Each choice involves a tradeoff between speed, quality, representation, and cost.

What do you think?

DISCUSSION QUESTIONS

• If an AI tool produces biased output, who bears responsibility — the developers, the training data, or the user who deploys the output?

• Is it enough to fix bias at the tool level, or does the problem require changing what gets created and published in the first place?

• Should AI tools be required to disclose the demographic composition of their training data?

• How should designers handle situations where AI tools are less capable of representing certain communities?

• Can efforts to 'de-bias' AI inadvertently flatten cultural differences into a homogenized idea of diversity?