OUR Artificial intelligence has already begun to “steal” jobs in the European Parliament.

Fast computer systems with digital comfort in “Eurocratic” jargon are replacing hundreds of translators working in the EU, downsizing one of the largest and oldest departments of Brussels’ multilingual institutions.

“And this may just be the beginning, as new AI tools can increasingly replace humans,” Politico comments. Translators are important cogs in the EU’s complex machine, as every official text must be translated into the 24 official EU languages. before it enters into force. A few years ago, only people performed this titanic task. But no more.

Data from the European Commission, the executive body of the EU, shows that the translation department declined by 17% over the past decade due to increased use of machine translation.. Over the same period, the percentage of outsourced translations has increased from 26% to 37%, and fewer translators are now being hired to cope with the increased workload.

Gone are the days when translators of EU institutions spent hours leafing through dusty dictionaries. The new technology saves time, admits “veteran” translator Markus Foti, head of the “machine translation” department at the Commission. “The time I would have spent in 1999 going to the local library to read the relevant paragraph and type it up can now be spent translating,” Foti says in an interview with Politico.

The translators insist that concerns about their work are exaggerated. Despite the fact that their work is adapting to artificial intelligence, the human factor remains important, they say. “The world is changing and translation cannot be ignored,” said Spyridon Pylos, a former Commissioner who oversaw the introduction of translation machines in 2013.

“Machine translation helps translators, but cannot replace them. You always need expert confirmation.

How Translation Became Digital

When new member states joined the Union in the early 2000s, the number of EU official languages ​​increased. increased to 24, and the Union has developed better tools to meet the rapidly growing translation needs.

The Commission was forced to abandon its old translation system because it was ill-suited to the languages ​​of the new Eastern European members. “It was a rule-based system, according to which you had to create dictionaries, define grammar rules and transformation rules. So it was very cumbersome and demanding,” explains Dieter Ramer, Head of Data at the Commission’s translation service.

Increase in work, layoffs, burnout

The workload for translators has increased from about 2 million pages in 2013 to 2.5 million in 2022, according to a spokesman for the EU executive.

The new engine entered service in 2013, but was replaced four years later with an even more sophisticated and efficient system. This uses an artificial neural network (a computational structure that mimics neurons) to predict the sequence of words.

These advanced tools are cost effective and allow fewer employees to translate more and more EU legislation.