1. DeepL has implemented a new language AI with a large language model (LLM) specialized in translation and text correction. 2. The AI is designed to reduce the risk of misinformation and provide more contextually accurate and human-like translations and text generation. 3. In blind tests with professional translators, DeepL's new AI outperformed ChatGPT-4, Google Translate, and Microsoft Translator, showing significant improvements over DeepL's previous models.