"Don't You Just Upload It to ChatGPT?"
Points and comments are a snapshot, not live.
A freelance translator argues that AI cannot replace human translation work despite public assumptions.
The author, a freelance translator in Ottawa, describes encountering a gym acquaintance who suggested uploading translation work to ChatGPT to save time. The author explains that while AI can produce grammatically correct output, human translators do more: they adapt content, localize it, research terminology, ensure consistency, and convey intent naturally. The author has experimented with AI tools since fall and uses them as secondary tools (for spell-checking, flagging rule violations, extracting terminology) but says everything requires double-checking. AI invents acronyms, forgets to translate sentences, ignores provided terminology, and misses context entirely. The author argues professionals should not be paid less for using tools, just as roofers aren't paid less for using hammers. The acquaintance, a government HR director, admitted she cannot use AI reliably at work.
What commenters are saying
Top comments split on the article's long-term relevance. Some argue the author's position is temporary; within 5-10 years, AI will likely outperform human translators and developers, citing already-visible shifts in software engineering where developers prefer AI over human advice and documentation. Others counter that current AI lacks reliable factual recall and reasoning depth, and that society's choices about AI adoption are not inevitable. A parallel debate emerges over whether market shrinkage in human translation quality is unfortunate (users suffer from poor machine translations) or positive (resources can be redirected). One commenter notes human translators already produce poor-quality work, and machine translation has raised the baseline for low-stakes use.