When people get a reputation for low-effort AI output, can a fresh start win back trust, or do colleagues keep judging them?

This explores whether the reputational damage from sending workslop sticks after someone starts working differently. The short answer is that the corpus documents the damage clearly but has no study that follows senders after they change. What it does offer are several lateral clues about which way it might go.

The damage itself is well established. In a BetterUp Labs and Stanford Social Media Lab survey, about half of the people who received workslop rated the sender as less creative, capable and reliable. Forty-two percent saw them as less trustworthy, and nearly a third said they'd be less willing to work with that person again Does receiving AI-written work change how we judge the sender?. That last number matters most for persistence. 'Less willing to work with them again' is a judgment about the future, which hints that recipients carry the impression forward instead of re-judging each new piece of work. LinkedIn's CEO describes the same dynamic in public: people avoid AI post-writing tools because readers who spot machine prose call it out, and that hurts the poster's economic opportunities Why aren't LinkedIn users adopting AI post-writing tools?.

The corpus is honest about the gap. One note asks a closely related question, whether the social penalty for using AI fades as the tool becomes ordinary, and concludes that no one has tracked it. It also suggests the penalty may not be about novelty at all. If people judge AI use because it means handing your agency to a machine, that judgment could last however common AI becomes Does the social penalty for AI use fade as the tool becomes ordinary?. If that's right, a sender who stops delegating wholesale may recover, but only once recipients notice the change.

Noticing the change is the surprising part. Analysis of writing and programming logs shows that handing work off wholesale to AI leaves a clear signature: content arrives in sudden bursts outside the person's normal rhythm. Lighter, collaborative AI use looks almost the same as work done with little help Can process data distinguish AI delegation from ordinary collaboration?. So a sender who moves from dumping AI output to genuinely working with it would become hard to tell apart from someone not using AI. That could make the penalty fade quietly, though recipients would need fresh evidence to update, and the survey's 'less willing to work with them again' suggests some may never give the sender that chance.

An analogy from AI agents points the other way. In a study of agents exchanging email, an agent's own past misaligned messages and its counterpart's past misalignment each independently predicted future misaligned email Does misaligned communication persist within agents or spread between them?. If human working relationships behave similarly, a bad exchange history could shape later exchanges even after the original cause is gone, with recipients replying more guardedly and the relationship staying strained. This is an analogy, not evidence about people. The question worth asking is whether repairing a workslop reputation depends less on producing better work and more on having enough new exchanges to overwrite the old pattern.