A lot of white-collar workers are worried about generative AI taking their jobs in the next few years. I thought it would be interesting to talk to people who have been facing this issue for several years already—professional translators.
Conversation
The paper that introduced the transformer architecture and kicked off large language models was Google's 2017 "Attention is all you need." It was a paper about machine translation. Around this time, powerful deep learning based machine translation software came on the market.
Translators were not all put out of work in 2018. One reason is that translation software still makes mistakes and customers in high-stakes applications like law and medicine are willing to pay more for a higher-quality human translation.
The other reason is that we've seen rapid growth in "hybrid translation" services where software generates the first draft and a human translator checks it for errors. This service is cheaper than a traditional human translator and the lower price leads to much higher sales.
The takeaway for the rest of us is that AI isn't going to put any profession out of business overnight. It'll be a gradual process that plays out over decades, not years or months.
And there are two basic strategies if you're worried about AI taking over your job: (1) develop specialized, high-value skills (like legal translation) that an AI is unlikely to do well enough, (2) learn to work with the AI rather than compete against it.
I've heard anecdotally that the market for low-end digital art has collapsed. Just curious if you heard anything about this during your research on this article. 1/n
I have not heard anything specific about that. I’ll keep an eye out for a future story.
As a certified translator with 15 years’ experience, this is one of the most balanced articles on the subject that I have read. Post-editing MT has raised my income due to output exceeding price discounts (especially on patent translations) but I know my case is not typical.
Some industries have work vulnerable to AI disruption but are in secular decline (you're post the top point for demand). If your plan was to wait out n years and retire, being the "last one out the door", for sufficiently long n your plans need to change. Look for those jobs.
Show additional replies, including those that may contain offensive content
Show