The convergence of AI and language technology is the culmination of a long history of losses and failures. 

What is so special about Machine Translation that we, humans, seem to never give up? There is something mysterious and religious about this business. In a conversation with the audience, Marco Trombetti and Jaap van der Meer, will dig up untold stories about the gurus and heroes who came before us, fighting the battle that we are still fighting today. Kicking off the first TAUS Massively Multilingual AI conference they will lay out a vision of what’s next, whether we like it or not.

Execs from AI Labs discuss the challenges and opportunities of massively multilingual models and applications. English is the dominant language in the world of tech and AI research. Models are trained for more than 90% on English language data. Inevitably the LLMs look at the world through English filters. The focus of this session will be how to overcome this language bias and how to move towards a more democratic and inclusive world of AI. Models and algorithms alone can not solve this problem on their own. In other words: AI needs the language industry, and the language industry needs AI. How convergence of two industries can accelerate our path towards truly multilingual AI.

Twelve innovators and disruptors have been selected to compete for the highest trust and scores from the audience presenting their AI solutions in six minutes.

Evolving the translation process at scale beyond just translation quality assessment or automatic post-editing. How can organizations leverage a coherent combination of human specialists, process management and new generation AI with both excitement and realistic pragmatism baked in. At a lower cost than yesterday

The unique capabilities of LLMs (large language models) create an unprecedented way to customize neural network behavior. LLMs are great for incorporating linguistic data and making machine translation customized with proper terms and phrases. memoQ AGT offers machine translation with instant domain adaptation accessible to anyone.

The presentation will introduce fly.qc to the audience. This innovative tool enables users to synchronize videos offline, facilitating efficient collaboration and review. With fly.qc, users can effortlessly add comments, generate detailed quality reports, and ensure consistent localization across all multimedia content. By streamlining quality control within the localization workflow, fly.qc significantly reduces errors, saves time, and enhances overall efficiency. Experience the power of fly.qc and elevate your video localization efforts to new heights.

Current LLM-based products and services are limited to single tasks (e.g., custom translation) and offer users minimal control. This contrasts with the potential of LLMs to handle various language tasks and be user-friendly for non-experts. PromptLoc democratizes AI solutions by allowing users to customize and control operations through natural language prompts. These prompts can include dynamic placeholders for specific details in xliff files, making the tool flexible for any application. It's like a Swiss army knife for localization tasks, enabling creative solutions and expanding accessibility.

In media localization workflows, consistent machine translation (MT) for speaker gender is crucial. AppTek's innovative approach integrates predictive speaker gender labels into their MT systems, reducing post-editing effort significantly. Deluxe's implementation showcases improved accuracy in gender-specific word forms across languages like Spanish and Hindi, alleviating gender bias while maintaining high overall MT quality. This presentation explores the seamless integration of predictive speaker labels and MT, promising enhanced efficiency and accuracy in media localization processes.

Our AI-driven content assessment tool redefines the assessment of content compliance across a broad spectrum, including DEI principles and various regulatory frameworks. Utilizing cutting-edge AI algorithms, this tool automates the validation process for inclusivity, legal integrity, and regulatory adherence, identifying gaps and offering actionable insights to empower businesses to navigate the complexities of global standards faster and more efficiently.

Discover Janus Worldwide's Smart LQA: Revolutionizing Linguistic Quality Management with AI. This innovative system tailors Linguistic Data Management, automating processes to optimize efficiency and reduce costs. With customizable quality assessment programs, Smart LQA ensures accuracy across various language pairs and domains, adapting check frequency based on stability. Its multidimensional evaluation combines Translation Quality Index with accuracy and fluency metrics, providing comprehensive insights for informed decision-making. Explore Smart LQA's adaptable reporting features, offering full visibility and detailed analyses across clients, vendors, and projects, ultimately streamlining quality control in linguistic services.

Even though machine translation has come a long way thanks to the recent advances in LLMs like ChatGPT, it is still not up to par with professional translators. When Hubspot approached us to automatically evaluate translations, we set out to work in collaboration with Custom.MT. We were not only able to build a good critic. To our surprise we also succeeded in making the system provide accurate suggestions on how to improve the translations. So, we decided to use this ‘critic’ to automatically make the translations better. This became our high quality translation system. We integrated it into Excel so it becomes really easy to use and to evaluate the translations.

Flow by Centific is a self-serve AI localization portal designed to empower localization stakeholders and customers. It allows them to easily onboard projects and execute AI-infused tasks directly and efficiently, eliminating the need for tech intermediaries. Flow’s integrated AI features enhance operational efficiency, enabling users to shape the platform to their specific needs. By onboarding their own models and creating custom AI features, users can significantly boost efficiency and streamline localization efforts with just a few clicks.