Why did freelance writers lose Upwork income right after ChatGPT launched — and who exactly took the hit?
Why did Upwork freelancers lose earnings after ChatGPT's release?
This explores what the collection says about why freelancers on Upwork, especially writers, lost work and income after ChatGPT launched in November 2022, and what that tells us about who AI actually displaces.
This explores why freelancers on Upwork, especially writers, lost work and pay once ChatGPT arrived. The short answer the corpus gives is substitution. Clients who used to post small, well-defined jobs could now get a passable version themselves. A study comparing freelancers in AI-exposed occupations with those in less-exposed ones found employment fell about 2% and monthly earnings about 5.2%, with writing hit hardest at first Did ChatGPT's release reduce freelance writing work and pay?. That fits a separate finding about how people actually use ChatGPT: work use centers on writing, not on looking things up How is ChatGPT actually being used by real people?. The tasks that left the freelance market are the tasks the tool is mostly used for.
The surprise is who got hurt. Lab experiments often find that AI helps weaker workers most and narrows the skill gap, so you might expect top-rated freelancers to be safe. The Upwork data shows the opposite. A strong track record and good reviews gave no protection, and the best-performing freelancers may have been hit disproportionately hard Does a strong track record protect freelancers from AI?. One plausible reading is that clients weren't choosing between a good and a bad freelancer. They were choosing between hiring anyone at all and doing it themselves with ChatGPT, and quality doesn't help much when the job itself goes away.
The same pattern shows up outside freelancing, which suggests it's a general effect rather than something peculiar to Upwork. After ChatGPT launched, Stack Overflow posts fell 16% compared with similar Q&A sites, growing to about 25% within six months. The drop was largest for programming languages ChatGPT knew best Did ChatGPT cause Stack Overflow posting to decline?. As with the freelancers, the content that disappeared wasn't just the low-quality stuff Did ChatGPT displace only low-quality Stack Overflow posts?. When a tool can answer a request directly, demand drops across the board, not only at the low end.
The contrast with salaried work is useful here. Danish administrative records show employers adopting chatbots widely, yet two years on, employees' earnings and hours had barely moved; what changed first was which tasks people did Does AI chatbot adoption change worker pay and hours?. Employees have a job that gets reorganized around the tool. Freelancers sell individual tasks, so when a task can be automated, the income goes with it right away. Another line of research points to a slower, longer-term harm: the freelance work that remains shifts from producing things to checking AI output. That cuts off the paid practice freelancers rely on to keep their skills sharp, and unlike employees, they get no mentorship to make up for it Does AI turn freelance work into validation instead of creation?.
One limit: the collection documents *that* the losses happened and *where* they were concentrated much better than it explains client decision-making directly. No study here interviews the clients who stopped posting jobs, so 'clients substituted ChatGPT for hiring' is the best-supported explanation rather than a directly observed one.
Sources 7 notes
A difference-in-differences study of Upwork freelancers found that occupations most exposed to generative AI experienced lower employment and earnings after ChatGPT's November 2022 release, with writing work showing the largest initial impact.
Analysis of representative ChatGPT conversations shows non-work use growing from 53% to over 70%, while work use centers on writing tasks rather than information retrieval, concentrating among educated professionals.
An Upwork study found no evidence that past performance or employment history moderated ChatGPT's negative effects on freelancer employment. The data even suggests top freelancers were hit disproportionately hard, contrary to experimental findings favoring low-ability workers.
A difference-in-differences model comparing Stack Overflow to four Q&A sites with lower ChatGPT exposure found weekly posts dropped 16% post-launch, growing to around 25% within six months. The effect was largest for programming languages with more training data in ChatGPT.
Vote scores on Stack Overflow showed no significant change after ChatGPT's release, suggesting the displaced content included high-quality posts, not merely duplicates or poor-quality material. However, this conclusion relies on votes as a proxy for quality without expert validation.
Show all 7 sources
Danish administrative records show employers adopted chatbots widely and workers took on new AI-related tasks within two years of ChatGPT's launch, yet earnings and hours remained stable within a 2% margin. Task restructuring preceded measurable wage or employment changes.
Research suggests generative AI reorganizes freelance labor away from skill-building task completion toward AI output validation. This shift cuts off the paid practice through which gig workers stay competitive, especially compared to salaried employees who receive mentorship and support.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- The Short-Term Effects of Generative Artificial Intelligence on Employment: Evidence from an Online Labor Market
- Are Large Language Models a Threat to Digital Public Goods? Evidence from Activity on Stack Overflow
- From Producing to Validating: How AI Is Deskilling Freelancers
- How People Use ChatGPT
- Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI
- Signaling in the Age of AI: Evidence from Cover Letters
- The state of enterprise AI
- LLM Targeted Underperformance Disproportionately Impacts Vulnerable Users