Roberto Fonts’ Post

SAFE AI created guidance for the use of AI in interpreting. It focuses on autonomy, safety and accountability for the end-user as well as transparency of the AI quality. #Interpretation #LanguageServices #Innovation

The temptation to use so-called “artificial intelligence” (really, human intelligence regurgitated artificially) for translation and interpreting has never been greater. Why not, because “it works”! But what does “it works” really mean? Watch out for the numerous pitfalls. It may be fast, it may be able to produce words by the second, it may always be available, but do you want speed and quantity, or quality for your event? In UN Today, a warning is sent as to data confidentiality, hard to spot inaccuracies, and the fact that “Large Language Models are rife with gender, racial, and anti-LGBTQI bias, according to a recent study done by UNESCO.” After weeks, months preparing your event, watch out for the call of the AI sirens, and opt for experienced humans who will tailor their skills to your event. AIIC Consultant Interpreters ✍️ info@cig-interpreters.com #interpreting #translation #eventprofs #MICE https://lnkd.in/gdAgFgSz

New data reveals a clear trend: experience with AI-driven services like spoken interpreting, signed interpreting, captioning, subtitling, and transcription correlates with more positive perceptions of AI's capabilities. While the increase in positivity is moderate, the pattern is consistent – the more people use AI services, the higher they rate their quality. This insight from our 'Perceptions on Automated Interpreting' study could signal a turning point in the acceptance and integration of AI in language services. Read the full report https://lnkd.in/gQSwP7xW These insights from the Interpreting SAFE-AI Task Force study help us understand expectations and areas for development in AI-powered language services. Join us for a live webinar on March 06 as Hélène shares insights from 'Perceptions on Automated Interpreting' study and discuss the future of automated language services. https://lnkd.in/dD9FKcaw The SAFE-AI study is sponsored by MasterWord Boostlingo LanguageLine Solutions Universal Language Service, Inc. Sorenson Communications  Certified Languages International PGLS Akorbi AMN Healthcare Cesco Linguistic Services THE LANGUAGE GROUP, LLC Translation Station, Inc. Cross-Cultural Communications, LLC National Council on Interpreting in Health Care (NCIHC) Lango #InterpretingTrust #AICommunication #AI #GenAI #Interpreting #Languages #InterpretingSAFEAI #SafeAI #MarketResearch #CSAResearch

This is one of the more fascinating finding from CSA Research's report on automated interpreting for the SAFE AI Task Force. Exposure to AI tools helps improve quality perceptions. #Interpretingservices #interpreters #languageaccess

Healthcare organizations using AI to generate or translate content need to be keenly aware of the cultural, gender, and racial bias in those systems. They can cause big problems for healthcare organizations such as compliance issues, lack of customer trust, and ongoing inequalities.  In our latest blog post in our Eliminating Bias in AI-powered Translation series, you can read about:  - The various impacts of bias in AI translations - What organizations should do now to combat bias - A glimpse into future solutions Read the entire blog here: https://lnkd.in/eivJtZrV

As AI continues to shape global communication, the importance of preserving linguistic diversity has never been more crucial. A recent article discusses the ongoing challenges and biases within AI technologies, particularly in their handling of languages beyond English. This highlights a silent revolution we are part of—one where technology must be guided to respect and amplify all voices, not just the dominant ones. For businesses aiming to genuinely connect with global audiences, this isn’t just about avoiding the awkwardness of poor translations or cringe-worthy movie voiceovers. It’s about understanding and investing in robust translation practices that honor and reflect the linguistic richness of their customers. Tools like translation memories are not merely technical assets; they are repositories of context, culture, and nuance that make translations not only accurate but also resonant. In an era where AI can both bridge and widen gaps, the role of trusted translation services becomes pivotal. Human translators, interpreters, content creators, and everyone in between are at the forefront, ensuring that when businesses speak, they do so in a language that is true to their identity and respectful of their audience’s cultural context. We must push for AI and technology to serve as bridges to mutual understanding, not barriers. As we integrate AI into our communication strategies, let’s commit to enhancing its linguistic dexterity and cultural sensitivity. Let’s make sure our technological advancements celebrate diversity, rather than diminish it. Read more about these challenges and potential solutions for multilingual AI: https://lnkd.in/eh3ziwnu #AI #Multilingualism #TranslationServices #GlobalCommunication #LinguisticDiversity #CulturalSensitivity #LiaisonMultilingual

Number of the Day: 200 Million 📍 This is the number of active ChatGPT users in 2024. This impressive figure underscores the growing reliance on AI in various sectors. OpenAI's website receives over 1.6 billion visits each month, indicating that more and more people are turning to AI for assistance in their daily tasks. ChatGPT reached 1 million users in less than 5 days. For comparison, Instagram achieved 1 million users in 75 days, while Spotify took 5 months. However, in an era where artificial intelligence is transforming industries, one area remains dependant on human expertise supremacy: linguistics. While AI exceeds at processing vast amounts of data and generating responses based on pure analysis, it lacks the emotional intelligence inherent in human communication. What are the key challenges of AI in language translation? — Poor contextual understanding, hence, further ambiguity in language; — Lack of domain-specific knowledge may lead to factial mistakes; — "Lost in translation" kind of occasions due to cultural adaptation; — Inability to interprete and handle complex linguistic structures, tone of voice and emotions; These challenges lead to AI still requiring human supervision when it comes to linguistics because language is more than just words—it’s the bridge that connects people, develops relationships, builds trust. How do you use AI in your work? #ai #languages #artificialintellingence #technology

I was excited to see so many AI-related sessions at this year's Language Show. John Worne, Chief Executive of the Chartered Institute of Linguists (CIOL), provided an insightful exploration of AI's implications for linguists. John recognised the transformative role of technology in the translation industry, from terminology tools to machine translation. However, the emergence of generative AI, particularly ChatGPT, has created unprecedented disruption and dialogue in the field. John posed the critical question of whether linguists are at a crossroads as AI impacts workflows, client expectations, and capabilities. While calls to pause AI development are notable, John advocated for guidance on responsible AI use to mitigate risks, as halting progress is impractical. John emphasised that AI encompasses diverse technologies with varying maturity levels, some potentially dangerous if mishandled. Linguists' expertise in language structures positions them to identify biases in large language models (LLMs), which often arise from training data that over-represents English. John pointed out that while many English speakers find AI's language capabilities impressive, performance deteriorates for other languages, especially Asian and African ones. This discrepancy can lead to greater social inequality, particularly in public service contexts requiring accuracy and equitable access. John explained that the probability-based approach of generative AI often leads to overconfidence, resulting in errors and hallucinations. He discussed 'model collapse,' where an AI's self-training can degrade quality over time, reinforcing inaccuracies. Ethical considerations were a significant focus. John questioned the trustworthiness of AI-developing companies and their safeguards. Despite concerns, he acknowledged generative AI's potential to enhance productivity, with CIOL incorporating it while emphasising the necessity of human oversight. John highlighted CIOL's AI resource bank as a valuable tool for linguists, helping them navigate AI-related information. He also pointed to challenges like the unpredictable nature of LLMs in translation and limitations of evaluation metrics. Experts agree that a human-AI partnership is the most effective approach. John touched on global perspectives, observing greater anxiety in English-speaking countries compared to European nations with more regulatory confidence, and enthusiasm in Asia potentially accelerating AI adoption. John concluded by reinforcing the importance of professional confidence, encouraging language professionals to stay informed, share examples of AI's limitations, and collaborate to promote ethical AI practices.

This article is a fascinating read. AI seems to be far off when translating from English into other languages. I keep saying, just like for other industries, bus specific for the language industry…a human in the loop is critical!!! #liaisonmultilingual #ai #tranalation

Geeta Pillai  27 Feb 2024 13:22 EST "The Advisory Group on AI and Sign Language Interpreting recently unveiled a comprehensive report, spotlighting the urgent need for regulatory measures to ensure the ethical use of artificial intelligence (AI) in automated sign language interpreting. This pivotal document, derived from a detailed survey of American Sign Language (ASL) interpreting service users, aims to establish a robust legal framework to safeguard both the social and technological integrity of AI-mediated interpreting Unpacking the Report's Findings The report meticulously balances the social implications and the technological advancements necessary for AIxAI's ethical application in sign language interpreting. It emphasizes the critical need for establishing clear standards and benchmarks to guarantee the technology's effectiveness and ethical usage. The Advisory Group's findings draw attention to the nuanced social dynamics and cultural norms inherent in interpersonal interactions, underscoring their significance in the development and deployment of AI technologies for ASL interpreting." metaglossia_mundus: https://lnkd.in/gBhRTMfz