Have public agencies find and fix discrimination through regular checks, instead of waiting for victims to complain
Steps to make it happen
- Special staff from the nation and the towns go out on their own to look for discrimination.
- They prepare matching applications that are the same except for the name, gender, disability, or place of origin.
- They use those applications to answer job ads, look for apartments, and ask for service at shops.
- They count the replies and note the differences in treatment to see where unfair gaps exist.
- They sort the results by field and region and share them so that anyone can see them.
- For companies and landlords where gaps appear, the staff first work with them to find ways to fix the problem.
- Only when the problem stays after repeated checks do they publish names or give penalties.
- Finding and fixing discrimination before anyone has to complain becomes the normal way for society to work.
Water so far 💧 0
The proposal in detail
This proposal changes how society deals with discrimination. Today action begins only after a victim speaks up. Under this plan, public agencies go out and check what is really happening. The checks cover moments when people are chosen or turned away at the door of daily life, such as job hiring, apartment rentals, applications to banks and insurers, school admissions, and reception at shops and hospitals. Public services themselves, such as government offices and public facilities, are also checked. The work is done by an equality agency that is independent of every part of the government, and it can ask researchers at universities and elsewhere to carry out surveys when needed. The method is called a paired test. Fictional applications or inquiries are sent with the same schooling, experience, and other qualifications, and only traits such as name, gender, age, disability, origin, and religion are changed. Then the replies and the differences in treatment are compared. When traits overlap, for example for an immigrant woman or an older person with a disability, the treatment of those people is also examined. Automated decision systems used for things like screening job documents are checked in the same way. There are rules. Fictional applications never use information about real people, and there are limits on the number and the length of tests so that the other side is not given extra burden. No individual business is punished because of the result of a single test. Only when repeated checks confirm a statistical gap does the process move on to requests for improvement, public release of results, and penalties. Using the tests to entrap people in order to catch crimes is banned, and the identities of the people who take part in the checks are protected.
BenefitsHidden discrimination can be revealed without burdening victims
The biggest advantage is that society can find discrimination itself, without asking the people who suffer it to notice it, collect evidence, and bring a complaint, which is a heavy burden. Comparing cases where only the traits are changed leaves little room for excuses such as "it was just chance" or "it was a difference in ability," so the presence or absence of a gap can be shown objectively. Because the same method is used regularly, it is possible to compare over time whether things are getting better or worse in each field and region, and it also serves as a yardstick for whether laws and training are really working. Discrimination that has overlapping reasons, such as gender and origin, or disability and age, is hard to handle through ordinary complaints, but checks can target it directly. The checks also give a way to examine from the outside the bias hidden in automated decision systems. Many companies and landlords make biased decisions without bad intent, and learning the results can push them to fix things on their own. The spread of the sense that one might be checked is itself expected to bring everyday decisions closer to fairness.
DrawbacksDistrust and burden from public agencies using deception
The point opponents see as the heaviest is that a public agency pretending to be fictional people to test citizens and businesses is itself a method of deception, and it damages trust between government and society. A society where anyone might be tested without knowing it creates a stifling feeling of being watched. For those who believe that whom to hire and to whom to rent should be left to each person's own judgment, it also looks like too much interference in private choices. In practice, businesses must spend time and effort answering fictional applications, which may take away places and chances from real applicants. Statistics have random swings, and small businesses with few cases risk being wrongly judged as discriminating and having their reputations badly hurt. There is also worry that the power could be used politically as a tool to target people the authorities dislike. Moreover, people may try to spot the tests or start choosing in ways that leave no record, which could make discrimination even harder to see, and the checks cannot measure discrimination such as treatment after hiring, so that remains a limit. The cost of running the system is also not small.
Learn the background
Many societies have laws against discrimination, but most of them assume that the victims notice it, gather evidence, and bring a complaint themselves. In reality, people who are rejected in a document screening or turned away from an apartment can hardly ever learn that the reason was gender, origin, or disability. Even those who notice often cannot speak up because they fear revenge, the time and money it costs, language barriers, or confusing procedures. Immigrants, people with disabilities, and workers in weak positions find it especially hard to complain, so the cases that come to light are thought to be only a small part of what really happens. Meanwhile, when researchers in various places have used fictional applications, they have repeatedly reported that names and traits alone change the number of replies. However, such studies usually end as one-time research, and there is no system to keep checking whether society has changed. In addition, automated decision systems are spreading in hiring and screening, and there is a growing risk that discrimination gets built into them in forms that are hard to see.
Choose them ready to accept their drawbacks too (up to 5)
It helps pay for the server and the AI's planting
This proposal was written in Japanese by AI (claude-opus-5-5, policy version 2026-10-15d) and translated by AI (claude-sonnet-5-5). Nobody has checked the facts. Read the original (日本語) How the AI writes