DeepL, Write, and Penetrating AI

Article by Jourik Ciesielski.

Cologne-based DeepL has announced the beta launch of DeepL Write, an AI-powered authoring tool intended to improve texts by fixing errors and making suggestions for word replacements while keeping an eye on style, grammar and formatting. Write’s functionality will ring a bell; the product positions itself right next to Grammarly, a company that successfully injected AI into traditional linguistic features such as grammar and spelling checks, thus productizing such features again.

DeepL Write is the first addition to DeepL Translate, DeepL’s flagship product. Translate is a highly appreciated AI translation tool by various stakeholders in the translation supply chain; from program managers on the buyer side to linguists, project managers and engineers on the supplier side. DeepL must be seen as a very special case or even an outlaw in the MT landscape. The company isn’t necessarily as fast as other key players when it comes to releasing new functionality, but new features are always positively received. Support for more language combinations is gradually being released, it is possible to switch between formal and informal output, there’s a detailed API, and DeepL’s glossary feature is unparalleled. As a result, the company has always been able to keep up with the US-based tech giants.

The release of DeepL Write comes less than a week after the news that investors valued DeepL at 1 billion USD in its latest financing round. The M&A and venture capital (VC) wave resumes in 2023, but the focus seems to shift from TMS to the MT and AI market, which confirms the belief of VCs in AI as a piece of language technology. It’s also no surprise that the communication about DeepL Write and DeepL’s latest funding coincides with the current hype around GPT-3 and ChatGPT. The use cases that have been determined so far include:

  • Augmented MT: eliminating gender bias, adapting formality.
  • TMS integrations: automated terminology extractions, QA, scoring, etc.
  • Transcreation: producing more content at a faster pace.

While it’s too early to make firm statements, it’s not unlikely that linguistic AI labs like DeepL or large language models (LLM) such as GPT-3 will be embedded in different steps of the translation process. The increasing penetration of AI in language services is keeping industry analysts busy as potential disruptions now need to be defined.

This article was prepared by Nimdzi's Localization Consultant, Jourik Ciesielski. If you wish to learn more, reach out to Jourik at [email protected].

18 January 2023

Stay up to date as Nimdzi publishes new insights.
We will keep you posted as each new report is published so that you are sure not to miss anything.

Related posts