jobs in the loop
What AI is doing to shifts, salaries, and the daily shape of a job — from the people whose work changed first.
Mara QuinnNarrator, Work and Money
Work
An Arizona caseworker was hired to help families apply for benefits. An English-first chatbot turned her into the person who repaired its Spanish intake summaries.
Mara Quinn · November 6, 2026 · 8 min
Work
After an injury, an overnight grocery stocker kept getting heavy work from an AI routing system. Refusing each task made his medical restriction harder to keep private.
Mara Quinn · 7 min read
Work
The cost plan looked finished, and executives had promised its total to a client. The estimator found that the AI had omitted fireproofing and accessibility work.
Mara Quinn · 7 min read
Work
A landscaping crew’s routing app cut miles by placing exposed properties together during the hottest part of the day. Their leader kept a notebook and built a safer route that scored worse.
Mara Quinn · 8 min read
Work
The AI earpiece saved time on visit notes. Then a home health aide found that it turned chest pressure and stopped medication into calmer, less urgent statements.
Mara Quinn · 7 min read
Work
A crew leader’s route kept getting denser during a heat advisory. Slowing down protected the workers, but the same system recorded them as inefficient.
Mara Quinn · 7 min read
Work
The tool gave a bedside nurse cleaner drafts and less typing. It also converted uncertainty into fact, leaving her to find every altered meaning before the chart became permanent.
Mara Quinn · 8 min read
Work
A home health aide found that an AI headset made urgent complaints sound mild. Checking its work protected patients but pushed her route further behind.
Mara Quinn · 7 min read
Work
The state-funded career tool matched Marcus to his old wage, but ignored his 25-pound lifting limit until he entered health information he did not want stored.
Mara Quinn · 8 min read
Work
The chatbot told a young apprentice how to isolate a wiring fault. Her notebook showed that its confident answer had ignored an energized circuit.
Mara Quinn · 8 min read
Work
Checkboxes from home visits became polished summaries that included meals, exercises and personal care that never happened. Correcting the record meant unpaid work.
Mara Quinn · 7 min read
Work
A tribal clinic’s automated phone system sent callers to the wrong records when it misheard local surnames. The receptionist spent 11 months documenting corrections without added pay or authority.
Mara Quinn · 7 min read
Work
Workers were warned for falling below 85 percent efficiency. Their union steward found that the AI benchmark changed while unscanned work counted against them.
Mara Quinn · 8 min read
Work
The worker’s score fell to 73 percent after approved training vanished from the denominator. A union steward found the change by comparing two printed reports.
Mara Quinn · 7 min read
Work
The tool cut routine paperwork nearly in half. On harder calls, its polished mistakes pushed a rural paramedic to document each scene twice.
Mara Quinn · 7 min read
Work
Customers rated the Florida call center agent highly. Voice software did not, and nobody supervising her could say which parts of her speech pulled the score down.
Mara Quinn · 7 min read
Work
The generated summary cut minutes from a crowded shift, but it left out the medication fact that mattered most. Checking the machine became another part of the nurse’s job.
Mara Quinn · 7 min read
Work
The generated handoff cut chart reading in half. Then a nurse found a penicillin allergy missing from the page she was expected to trust.
Mara Quinn · 7 min read
Work
A collections assistant was hired to bring museum records into the present. A chatbot kept writing about her living community in the past tense.
Mara Quinn · 8 min read
Work
Her title and $68,000 salary stayed the same. A spreadsheet showed that reviewing machine-written copy meant more output, more corrections, and less room to write.
Mara Quinn · 7 min read
Work
A live court transcript changed “ex-wife” to “wife.” Catching the error meant an interpreter had to split her attention during testimony.
Mara Quinn · 7 min read
Work
An experienced paramedic kept a notebook of generated reports that made uncertain observations sound final. Fixing them delayed the crew; leaving them could affect care and later reviews.
Mara Quinn · 8 min read
Work
The map showed that fire had crossed an access road. The machine-written briefing still described the road as outside the perimeter, leaving a seasonal firefighter to press for a pause.
Mara Quinn · 7 min read
Work
After four job applications were rejected in under a minute, a warehouse supervisor changed his résumé and tracked what passed the screen.
Mara Quinn · 8 min read
One story a day
One real story about AI each morning — no hype, no alarm, just company for the road.
