AI tools are changing translation work and cutting demand for humans

In short: AI translation is turning many translation jobs into faster, lower-paid editing work, and the number of human translators is falling.

What's going on

Professional translators say their work has changed since machine translation tools became common. Instead of translating from scratch, many are now asked to do “machine translation post-editing,” which means fixing and polishing text that a computer translated first.

Several translators told the Financial Times that agencies expect this editing to be done much faster than full translation, so they pay less. One subtitle translator said a major agency cut pay from about $5 per minute of video to $1.50 per minute after switching to post-editing. Another translator working on legal and academic text said rates per word fell by about 50 percent.

Translators also said the new workflow can be mentally tiring. To check a machine translation properly, you often have to compare two texts at once, the original and the machine version, like proofreading a document while also fact-checking it line by line.

The Financial Times also points to US labor data showing a longer-term shift. Growth in the number of translators slowed around 2010, around the time Google Translate became widely used, and pay compared with the overall economy began to fall. In the last five years, the share of translators in US employment has dropped by almost 20 percent from its peak, and typical wages have fallen below the whole-economy average.

What to watch

Some experienced translators say they refuse post-editing work, which may push more work to less experienced workers. The article also suggests more companies may skip human editing entirely as they become more comfortable with fully automated translations, which could further reduce freelance opportunities.

Source: Financial Times