This past week, several Smartling team members participated in the 2024 GALA Conference held in Valencia, Spain, marking its largest GALA gathering to date. The conference theme, "Navigating the Rise of AI: Dos and Don'ts, Myths and Truths for a New AI-Assisted Localization Industry," deeply resonated with our mission at Smartling.
Our focus extends beyond simply implementing AI, following the technology hype; we strive to help our customers envision comprehensive applications of AI across their translation process. We found immense value in hearing insights from industry leaders during this event and are eager to integrate these perspectives into our ongoing research efforts.
In case you missed the event, we've distilled some key insights from our team, along with a summary of the perspectives Smartling shared during the event.
Smartling Team Takeaways
- A collective understanding of AI’s impact: As an industry, there is a shared recognition of imminent change as AI begins to replace common tasks. While this shift isn’t unprecedented - think back to the introduction of first CAT tools, then the three generations of machine translation - there’s a shared goal to grasp how our field will evolve amidst these changes. Adaptation to new paradigms, challenges, business engagements and modes of thought will be crucial.
- The quest for AI’s tangible value: People throughout the industry are still looking for an answer to the question: how can AI be applied and provide real value within the industry? Debates and discussions on this topic were heard consistently during the conference, tinged with apprehension about AI's implications and ethical considerations on its use in various scenarios.
- The search for leadership: In this AI landscape, the industry is still looking for a leader to help guide the true universal use of AI in our every work. While many companies are looking into specific, AI-related issues in translation, a unifying figure to champion best practices remains elusive.
- The ethical and regulatory impact of AI: Individuals in the industry are questioning how we can ensure we both embrace the unlimited opportunities that AI presents but also mitigate the rising concerns about the way AI applications handle personal information, introduce cultural biases, and present potential risks for data leaks and copyright infringement. Through collaborative discussions, enterprise clients, technology providers, and academic stakeholders have engaged deeply in reflecting on the ethical implications of integrating AI into the fabric of the localization industry. They've also pondered the transformative effects AI will have on job dynamics and the requisite skill sets of our workforce.
Sharing our perspectives
Smartling had the honor of presenting on several topics over the two day conference, offering insights spanning from AI Ethics to pioneering software paradigms and forecasting the emergence of Language Co-pilots. Here’s a summary of what we shared:
Ethics of AI in the Translation Industry - Aren’t we the most impacted field?
Our VP of AI, Olga Beregovaya, hosted a panel discussion featuring industry experts delving into the ethical dimensions of AI development and utilization within localization. Joined by Jose Palomares from Coupa Software, Manuel Herranz from Pangeanic, and Mrinalini Kochupillai from the Technical University of Munich, we explored how to address ethics and “responsible AI” as an industry and what we can do about biased and/or proprietary data that is used to train LLMs and ML models.
As primary contributors to global communication across diverse industries, we bear dual responsibility: for the ethical use of data and for shaping communication in every language. Moreover, it's imperative that we cultivate an environment where industry professionals can thrive, succeed, and find fulfillment amidst the rise of AI.
During the discussions, the panelists emphasized the significance of the emerging role of a "corporate ethicist," now becoming a staple in enterprises worldwide. This role is crucial for comprehensively understanding, assessing, and addressing the ethical implications inherent in AI implementation.
Mrinalini Kochupillai, offering insights from academia beyond the translation sector, provided a broader perspective on the social responsibility and ethical considerations surrounding AI applications across various industries and verticals. Her anecdote about AI assistants in the medical field underscored the importance of deploying AI judiciously, particularly in contexts where it aligns most effectively, and exercising caution in hastily implemented solutions, such as those involving mental health patients.
Death of the TM - or Long Live the TM?
Olga also presented a re-evaluation of the future of technology providers in the localization industry. She advocated for a reconceptualization of the translation landscape, urging a shift beyond the conventional reliance solely on Translation Memory (TM) and established industry standards, such as XLIFF and TMX.
Instead, she proposed transitioning from the traditional one-to-one, source-to-translation framework to a dynamic many-to-many relationship model, and possibly dynamically updated ontologies, capturing content concepts and relationships between them. In this envisioned future, TMs will evolve to assume a different role—not merely as repositories for direct matching, but as reservoirs of existing communication to be accessed and utilized by generative AI during translation or content generation processes.
Emergence of Language Co-Pilots.Jennifer Wong, Head of AI Product Strategy, challenged the industry to proactively envision and develop new tools and skills necessary to equip linguists and localization managers for the imminent integration of AI into localization processes. As Smartling investigates how to enhance linguist tooling, we also recognize a pressing need, and growing interest from the educational sector, to prepare new linguists and localization professionals for an evolving landscape where curiosity and adaptability are the key skills.
LLM, the multi-thousand dollar solution for the ten-dollar problem.
Jennifer also presented a session where she questioned the industry’s approach of rapidly rolling out AI solutions without sufficient scrutiny. With AI being such a wave maker, are we really solving the problems we should be solving with AI? Many conference attendees echoed similar sentiments, emphasizing the importance of assessing the costs, both monetary and environmental, associated with deploying AI for non-essential solutions. This includes considerations such as the sustainability implications of utilizing larger models, fine-tuning, and ongoing model training.
At Smartling, while we recognize the significance of AI in our industry both presently and in the future, we advocate for a balanced approach that prioritizes understanding the real Return on Investment for the features we develop and the research we undertake. We are committed to delivering quality features derived from thorough research and validation, ensuring that our solutions benefit not only our customers but also our planet.
Olga also served on the advisory committee for the conference. Together with industry peers from various industry segments, she helped shape the conference theme and ensure the highest quality, relevant content.
Coming back from such an impactful conference has really inspired the team. We are looking forward to sharing more of our thoughts and learnings at our upcoming Global Ready Conference. Stay tuned for more about how Smartling is working to bring LanguageAI to our customers in a holistic way.
