Authorship dilution: why AI-assisted rejection may feel different.

Authorship dilution is our name for the weaker bond between personal effort and an outcome when a tool performs part of the production work. In a job search, an agent can carry repeated form work. A rejection is then tied to less manual labor, even though the role and the decision still matter.

The term is a product idea, not a diagnosis or an established effect in psychology. No study in the sources below proves that an AI-assisted rejection hurts less. The research supports a narrower claim. Effort, generation, ownership, and reliance can change how people relate to what they make.

That distinction matters. A job application is not a flat-pack table, and a rejection is not a product rating. A person can feel rejected because the employer declined their experience, identity, hopes, or need for work. Software should not turn that pain into a slogan or promise emotional immunity.

The useful question is practical. Which parts of an application need the candidate's authorship, and which parts only consume the candidate's time? Jobwright is designed to keep the first and reduce the second.

The IKEA effect describes a pattern in which people value something more after they help make it. Michael Norton, Daniel Mochon, and Dan Ariely found that successful assembly could raise the value participants placed on ordinary products. The result depended on completion. Labor alone was not magic.

Effort justification points in a related direction. In a classic 1959 experiment, people who went through a difficult initiation rated a dull group more favorably than people who entered with little effort. The setting was a group experiment, not hiring. It still shows why effort can alter the meaning assigned to an outcome.

Psychological ownership research gives the broader idea a useful name. People can experience something as mine without legal ownership. Control, close knowledge, and personal investment can help produce that feeling. A carefully written application can therefore become more than a document. It can feel like a small container for the applicant's judgment and hope.

That attachment has value. It can produce care, accuracy, and a stronger account of the work. It also adds emotional cost when the same care is repeated across fifty forms that never receive a human reply.

Authorship dilution targets that repeated cost. The agent does not make the candidate less real. It makes the form a smaller claim on the candidate's time and craft.

The generation effect is a warning against handing the whole story to a model. Norman Slamecka and Peter Graf found that people remembered generated words better than words they only read. Later work has tested the effect in many learning settings. The original finding concerns memory, not resumes, but the design lesson is sensible.

A candidate should generate the core examples that explain their work. They should decide which result mattered, what trade-off they made, and which detail they can defend in an interview. If the model invents that material, the application may be fluent while the candidate has no firm memory of how the story was built.

Cognitive offloading is useful when it reduces demand without removing the knowledge needed for the task. A calendar can remember an appointment while the person still decides whether to attend. In the same way, an agent can transfer facts into a form while the candidate remains responsible for the facts and the reason for applying.

The right goal is not maximum delegation. It is selective authorship. Generate the parts that form identity and judgment. Offload the parts that repeat a decision already made.

Craft alienation is a descriptive phrase here, not a precise experimental term. It names the feeling of becoming separated from the process and product of your own work. An application can create that feeling when the prose sounds polished but no longer sounds like the person who must stand behind it.

This failure can happen without a factual lie. A model may flatten a specific example into corporate language. It may turn uncertainty into confidence, remove the detail that made the work memorable, or describe a motivation the candidate would never say aloud. The form remains plausible, but authorship has moved too far away.

That distance creates operational risk. The candidate may struggle to answer a follow-up question. A recruiter may hear a different voice in the interview. The agent may reuse a sentence after the facts or target role have changed. Fluency hides the break until a person asks for detail.

Jobwright cannot solve this by adding a generic instruction to be authentic. It has to bind claims to verified candidate facts, show the evidence used for a revision, and stop when an answer is unknown or sensitive. The interface must make correction easier than passive acceptance.

A useful split follows the source of authority. The candidate authors the facts, target, exclusions, selected documents, sensitive answers, and campaign limits. The agent can research a role, compare requirements, draft a grounded revision, stage routine fields, and collect evidence for review.

Review mode keeps an exact approval before final submission. Search & Apply uses a different form of authorship. The candidate defines and accepts a bounded campaign mandate. At execution time, the system must still prove that the exact task, destination, role, data, document, and action fit that mandate. A mismatch returns the action to Review.

Research on trust in automation argues for appropriate reliance rather than blind trust or blanket refusal. That is the right frame for an application agent. The interface should expose enough evidence for a person to know when to rely, when to inspect, and when to take control.

This boundary preserves the candidate's authorship of consequence. It dilutes only the production labor that can be repeated from verified choices.

• You author the career facts, examples, goals, limits, and permissions.
• The agent handles repeated search, comparison, drafting, staging, and evidence assembly.
• The state service and browser gate check authority again when a protected action runs.
• Unknown, sensitive, or out-of-scope work returns to you.

Authorship dilution is not emotional armor. A rejection may still carry money pressure, lost hope, identity threat, or simple disappointment. A system that treats the candidate as a conversion metric would deepen the alienation it claims to reduce.

The narrower promise is about labor. A candidate should not have to rebuild the same factual packet, rewrite the same history, and move the same answers through another brittle form before an employer can say no. An agent can carry that repetition while keeping the candidate's claims inspectable.

This changes the object that receives the rejection. The employer still declines the application, but the application contains fewer hours of hand-built form work. The candidate keeps more time for choosing roles, improving the underlying story, preparing for interviews, resting, or leaving a bad search alone for a day.

That is why the catchphrase needs two parts. Delegate the form. Keep the voice. Dilution should reduce the bond between repetitive labor and outcome. It should not dilute truth, memory, consent, or the ability to say, yes, this application represents me.

A rejection can remain personal without becoming a verdict on an evening spent copying one life into another form.