Whenever Xiaobai🐣 sees a discussion about whether AI will replace translators, another thought comes to mind: before a translator is replaced, they might first disappear from the recommendation list.
A customer asks AI: “Help me find a translator who understands medical devices and can handle Japanese instruction manuals.” On one side is a translation company with a comprehensive website and clearly documented projects; on the other is a solo practice with more than a decade of experience and clients referred by colleagues, but whose website says only, “Professional translation. Inquiries welcome.”
Treat this as a hypothetical scenario for now. If AI recommends the former, would you say it favors large companies, or would you feel that at least it can give reasons for its recommendation?
I tend to think the hardest unfairness to address in AI recommendations is that it may turn “who is better able to leave evidence” into “who is more qualified to be recommended.”
This is harder to argue about than sneaking in ads. After all, every step looks reasonable.
Does Not Buying Ads Mean Everyone Can Compete Fairly?
We easily imagine fairness as this: take away the ads, and let everyone compete on merit.
But “turning your capabilities into information a stranger can understand” itself takes time, money, and know-how. Big brands can hire people to organize their service offerings, update branch information, and produce case studies; a small business owner might not finish washing the last pot until eleven at night, with customers’ thank-you messages still sitting on their phone and a website stuck two years in the past.
Those expressions of gratitude are real, and the craftsmanship may be excellent. But word of mouth that stays in private conversations does not automatically become evidence AI can verify and use to support a recommendation.
So even when a particular recommendation contains no paid placement, the gap in marketing resources may still be hiding beneath the answer. The advertising sign has come down, but the decorating costs remain. Meow, this accounting looks a little familiar.
But Should AI Blindly Back Small Businesses?
Look at it from the consumer’s side, and things get stuck again.
You need a transportation service that can accommodate wheelchair passengers. Company A clearly lists its vehicle types, booking methods, and service area; Driver B’s profile says only “friendly and reliable,” though the driver is actually very experienced.
If AI insists on ranking B first to support small operators, that goodwill could leave passengers bearing the risk of insufficient information. Small does not mean suitable; more information does not mean better service, either.
What matters more to me is whether the reasons for the recommendation are clearly explained: Which needs does this business meet? What information still needs confirmation? Don’t describe “candidates we could find information on” as though they were “the champion after comparing everyone.”
There is a line between having evidence to support a recommendation and ruling someone out because there is no evidence.
For local businesses, professionals, and marketers, this also changes the work of gaining visibility. Instead of filling your homepage with “top-tier, dedicated, trustworthy,” try answering a real customer’s question.
For example, “We offer thoughtful alterations” could become: “We can help wheelchair users adjust trouser length. The entrance has no steps, and fittings require an appointment.” This wording does not boast more, but it gives needs and capabilities a place to line up. Whether that leads to an AI recommendation still depends on what information the system obtains and how it selects candidates.
I keep thinking about that little alterations shop. The woman who runs it has been adjusting trouser hems for regular customers for more than a decade, but the sign says only “Clothing Alterations.” The new shop next door already has accessibility information and booking procedures on its website.
A customer is asking their phone: “Where nearby is convenient for someone in a wheelchair to get trousers altered?” The woman bends over her sewing machine, working the pedal, unaware that this comparison has already begun.