DeepL, the AI-driven translation company that is increasingly working with law firms – including Taylor Wessing – has announced updates to its API, including the integration of their ‘DeepL next-gen language model’ and its ‘Write’ capability for drafting, which includes working with contracts.
The company, which is valued at $2 billion after significant funding rounds and growth, said that: ‘The next-gen model delivers more precise, accurate translations; while the introduction of the DeepL API for Write [its drafting tool] provides users access to the company’s advanced writing tools, including grammar suggestions, spelling corrections and more.’
They also stated that their new translation capability is empirically better than previously and other competitors, such as Google and OpenAI.
‘Blind tests conducted with leading linguists found a remarkable 1.7x improvement with the new LLM against DeepL’s old model for combinations involving English to Japanese and Simplified Chinese; and a 1.4x improvement for the combination of English and German.
‘DeepL’s next-gen model is also proven to significantly outperform competitors in translation quality: in 2024, blind tests found that DeepL’s translations require fewer edits than competitors, with Google Translate needing 2x more edits and ChatGPT-4 needing 3x more edits to achieve the same quality,’ they claimed.
Meanwhile, in terms of helping lawyers – and others – to draft, its updated Write capability can now: ‘Go above and beyond what is offered by common generative AI tools that auto-populate text or basic grammar correction tools.’ And it also can leverage an API.
‘DeepL Write serves as a creative writing assistant, providing real-time, AI-powered suggestions on word choice, phrasing, style and tone during the drafting process. The addition of this tool will transform the way knowledge workers create business content – improving the accuracy and quality of everything from internal communications to customer-facing messages and contracts on a global scale.’
I.e. The legal sector is very much part of its target client base.
Sebastian Enderlein, CTO, DeepL, concluded: ‘With the DeepL API, businesses can integrate our Language AI tools directly into their own internal and external applications, such as email workflows, e-commerce platforms and more, to overcome challenges ranging from language barriers, to content localization and writing clarity. It also maintains the rigorous enterprise-grade security standards that we are known for, so that sensitive data is always protected.’