“The Impact of Instant Translation is Profound”: Skrivanek’s Arturs Peha on How AI is Changing Language

Why Trust Techopedia

In the fast-paced world of language services, the fusion of artificial intelligence (AI) and machine translation powered by artificial neural networks is transforming the field, bringing efficiency and accuracy to translation services.

Guided by Arturs Peha, the Chief Strategy Officer (CSO), Skrivanek is at the center of pushing the dials on translation and localization solutions.

In an engaging conversation with Arturs Peha, we dive deep into how Skrivanek blends AI into its operations, the role of NMT in enhancing translation services, and what the future holds for AI and NMT in this industry.

Through Arturs’ lens, we will uncover the key advantages, challenges, and moral questions surrounding the integration of these cutting-edge technologies, offering a vivid snapshot of AI’s transformative power in translation today and its potential for tomorrow.

Key Takeaways

  • The integration of AI and machine translation is reshaping language services, both in the business world and on a human level.
  • However, while AI tools offer advantages, human insight, and oversight remain crucial for addressing subtleties and cultural nuances.
  • Ethically, we need to consider privacy protection, bias mitigation, and preserving cultural diversity.
  • Over the next five years, AI and NMT will become more commonplace, impacting job markets, business operations, and societal norms.

About Arturs Peha

Arturs Peha

Arturs Peha holds the position of Chief Strategy Officer at Skrivanek, where he has played a crucial role in merging technology with language services. His journey at Skrivanek, from leading development teams to shaping strategic directions, highlights his commitment to leveraging technological innovations in the translation industry. His work, especially on projects like the Skrivanek NMT engine and the Language Knowledge Assessment Tool, makes Skrivanek a front-runner status in adopting AI and NMT solutions.

His deep involvement in technology deployment and strategy within Skrivanek makes him a rich source of insights into the evolution and future of AI and NMT in the translation sector.


Human Role and AI Impact on the Translation Industry

Q: How does the advancement of AI and NMT technologies influence the demand for human translators?

A: The development of AI and NMT technologies has changed the kind and amount of work that specialists are expected to do. This doesn’t always mean they earn more, but it does lead to new kinds of projects for those providing localization services.

AI tools boost efficiency and take over routine tasks, yet the insight of humans is vital for addressing the subtleties, cultural nuances, and specific areas of knowledge.

At Skrivanek, human translators are essential for checking, correcting, and assessing AI-generated content, making sure the end results are accurate and culturally relevant.

Like in other sectors—whether marketing, legal, or IT—there’s growing discussion around the post-editing era, mirroring a conversation that began among translators some time ago.

Many feel quite threatened by the rise of AI, and rightly so, as it suggests that not everyone will keep their place in the industry.

This shift is prompting individuals who once leaned heavily on freelance work to seek more permanent roles in the job market, aiming for a sense of stability amidst the changes.

Real-world Impacts and Future Directions

Q: Can you discuss any recent advancements in AI that have particularly excited or concerned you in terms of translation?

A: The words “threat” and “chance” have six letters, so we can observe advancements from both perspectives, but I choose to view them from the “opportunity” perspective.

To me, the opportunity to speak in a mixed language seems intriguing, such as partially using some words in Polish and some in English, with a machine interpreting solution being able to comprehend these mixed languages and still provide transcription/instant translation in the desired language.

Let’s be frank: language service providers address only a tiny fraction of global communication needs. What concerns me is that, with all these “instant solutions,” we, as a society, are becoming increasingly tolerant of “low-quality language”. This acceptance could have long-term consequences, especially for smaller languages, potentially undermining their complexity and richness.

Q: How do you generally see the real-world impacts of better translation across the world?

A: Generally, as a member of society and the global community, I view the improvements in translation technology as a blessing.

It enables me to communicate with people around the globe in languages I don’t know and most likely will never learn.

Does this mean we will understand each other as well as two native speakers do? Certainly not. A lot will be lost, especially cultural nuances. Is encountering problems by misusing self-service technologies, like choosing the wrong tool for a specific task is possible? Yes, this can lead to issues in both professional and personal contexts.

Despite advancements in technology, we have seen numerous cases where businesspeople reach out to us seeking support when AI/NMT [neural machine translation] did not meet their expectations in terms of marketing needs, etc.

Q: Does instant translation change business / even relationships?

A: Yes, I believe it does.

As mentioned earlier, in personal matters, instant translation allows you/me to at least attempt to be understood or to comprehend elements I would not have been able to grasp in the past without carrying around a massive dictionary.

Now, many of us enjoy e-commerce services, often facilitated by machine translation. In business, using similar technology is helpful, but the risks are greater. An error in a cold email or marketing campaign could lead to serious financial impacts.

Moreover, instant translation is transforming educational opportunities, enabling access to knowledge that was previously language-restricted. It can also affect tourism, where travelers can interact more deeply with local cultures and experiences, enhancing mutual respect and appreciation.

However, there’s a delicate balance to maintain between using technology to bridge language gaps and preserving the depth and nuances of personal and cultural interactions.

The impact of instant translation is profound, extending far beyond convenience to shaping how we connect, conduct business, and build relationships across the globe.

Q: How do you see AI and NMT evolving in the next five years, and what impact will this have on the translation industry?

A: Eventually, using AI for specific tasks will become commonplace. This will require people to adjust and find ways to benefit from these technologies.

As usual, large corporations will support/push their teams toward adoption. Currently, Small and Medium Enterprises (SMEs) have a huge advantage — they can adopt technology much faster, with less fear of loss, especially as technology becomes more accessible through pay-as-you-go models.

Additionally, we’ll see many emerging millionaires who figure out how to deploy Large Language Models (LLMs) and fine-tune them to solve micro problems. Governments will need to act quickly on policies such as minimum income, change taxation policies, and figure out how to tax robot usage.

This era will introduce numerous new jobs and make it easier for people to produce their own digital content, like games and videos. However, this could also lead to negative outcomes such as increased misinformation, radicalization, and challenges in protecting privacy and financial information.

As for the impact of AI and Generative AI across different sectors, it will largely depend on each country’s economy and societal openness:

  • In the translation and language services sector, we’ll see a rise in providers offering mixed services.
  • Some IT companies and call centers are already adding translation as a secondary service.
  • Language service providers are expanding into IT-related services, including system testing, setup, and social media management.
  • PR firms are broadening their service range as well.
  • The use of neural machine interpreting will grow in the field of interpreting, pushing society to accept a standard of “average quality” as the norm for speed and convenience in services.

Without adaptation, many businesses may fail, leading to more mergers and acquisitions, with smaller language service providers dropping in value and larger entities emerging, some new to the industry.

Integration and Advancement of AI in Translation

Q: Can you describe how Skrivanek integrates AI into its translation services?

A: We saw the potential for using AI in our work processes, even when only statistical machine translation systems were common. Our Research and Development (R&D) team is focused on creating unique tools and improving current AI technologies, paying special attention to adopting new technologies early to meet the unique requirements of the language industry.

We developed the Skrivanek NMT engine, a specially-made tool designed to cater to the distinct needs of clients in fields like law, medicine, and technology. We’ve also successfully created our Language Knowledge Assessment Tool, which is now used by many respected IT companies and HR departments.

Presently, our team is working on several projects related to the use of generative AI in everyday tasks, aiming at either enhancing our internal processes or creating new services that our clients will find valuable.

We also have teams ready to assist our clients with editing and quickly reviewing AI-generated content, including text, videos, sound, or graphics, regardless of whether the content was produced with our assistance or by the clients themselves.

Q: How does the use of NMT enhance the quality and efficiency of translation services compared to traditional methods?

A: The use of Neural Machine Translation (NMT) significantly increases the speed of work for translation and localization professionals. It automates repetitive tasks, thereby reducing the time and effort required for translation, and leads to the ability to lower the costs of translation, making machine translation more appealing.

In our experience at Skrivanek, we once assisted an e-commerce company looking to launch a product quickly. To achieve this, we initially used raw machine translation to get their content online swiftly, enabling search engines to start indexing it right away. This step was critical for the product’s immediate visibility. Shortly after, we refined this content through post-editing to ensure it met high-quality standards and accurately reflected the company’s branding.

In another instance, there was a project of 4 million words where the client had a very limited budget, and the deadline was just a few days away. It was finally proposed to manage the entire project using NMT, and the client was extremely satisfied with the end result as it matched their specific business needs.

Did we personally think that the quality of this machine translation was flawless? Of course not. But it clearly corresponded to client expectations.

What is great about using modern NMT and project management solutions is that they also include built-in quality assurance (QA) tools that further support the business’s needs by providing agile, affordable, and high-quality language solutions.

However, not all projects are suitable for NMT. Due to specific agreements like those with military clients or others with strict policies against machine translation, some require alternative approaches.

Overall, the use of NMT and related technologies is reshaping the landscape of translation, focusing on efficiency and accessibility while addressing the diverse needs of various clients and projects.

Benefits and Challenges of AI in Translation

Q: In your view, what are the most significant benefits of NMT for businesses today?

A: The most significant benefits of NMT for businesses today include increased productivity, enhanced speed of translation projects (including instant translation options), and the ability for language specialists to work more independently.

AI tools, including NMT, allow human translators to develop creative skills by focusing on nuances, cultural context, and idiomatic expressions.

This, in turn, enables businesses to adapt marketing materials for new audiences with greater creativity and cultural understanding.

Q: What are the main challenges you face when integrating AI into the translation process, and how do you overcome them?

A: Integrating AI into the translation process comes with its own set of challenges, though they are not overwhelming.

  • A significant concern for project managers is not just adding AI into their workflow but determining if a project is suitable for AI or if it matches the client’s needs. Additionally, there’s a noticeable shortage of skilled post-editors, particularly for certain language pairs where finding experts for thorough post-editing can be difficult.
  • There are also clients with very strict policies that limit companies from applying the post-editing process. Sometimes, there are obvious and logical reasons for this, but at times, it might be just connected to past bad experiences clients had.
  • A larger challenge for language service providers lies in adapting to and incorporating GenAI technologies into their everyday operations. This topic is of interest across various industries and is particularly relevant in the language services industry, which has long been shaped by technological progress and automation.

Q: What limitations have you encountered with current machine translation technologies, and how do you address them?

A: Modern machine translation tools face challenges in correctly translating specialized terms, managing idiomatic expressions, and grasping cultural nuances. To tackle these issues, we integrate these tools with computer-assisted translation software during our workflow. Human translators, who are experts in specific fields, then review and refine the machine-generated translations to ensure they are accurate and contextually relevant.

Adding more training data, including detailed terminology databases tailored to particular industries or areas, can enhance the machine translation systems’ ability to deal with complex terminology.

Conducting regular quality checks and establishing feedback loops are crucial steps in identifying and correcting any errors or shortcomings in the machine translation results.

Q: How does Skrivanek ensure the accuracy and cultural relevance of translations produced by AI technologies?

A: Skrivanek ensures the accuracy and cultural relevance of translations produced by AI technologies by combining machine translation tools, such as its own NMT engine, Google Translate, DeepL, Bing, and other neural machine translation engines, with human expertise.

Human vendors also work with generative AI software like ChatGPT to clarify the context of source fragments, summarize content, adjust target documents to fit formal requirements, and change the tone and style of the created content. This collaborative approach ensures that translations maintain accuracy and cultural relevance.

Ethical Considerations and Data Security

Q: What ethical considerations should be taken into account when implementing AI in translation services?

A: When considering the ethics of using AI and NMT, several key points stand out:

  • Privacy and Data Protection: We often work with very sensitive information, such as medical records, company secrets, or banking details, alongside more general marketing data. It’s crucial to have clear guidelines that outline when and how AI/NMT can be used safely and appropriately.
  • Bias in AI Systems: AI systems learn from data created by humans, who naturally have biases and prejudices. It’s vital to ensure these biases do not influence the outcomes of AI/NMT, leading to harmful results.
  • Impact on Lesser-Used Languages: Many NMT solutions use “bridge languages” to translate between less common languages. This could lead to smaller languages losing their unique cultural features and becoming more uniform.
  • Other Concerns: Issues such as the displacement of human workers, the potential for misusing these technologies, and disputes over intellectual property rights are increasingly relevant and require careful consideration.

These ethical considerations highlight the need for thoughtful and responsible use of AI and NMT technologies, balancing innovation with the preservation of privacy, cultural diversity, and fairness.

Q: How do you ensure the confidentiality and security of data when using AI and NMT for translation projects?

A: To protect the confidentiality and security of data in AI and NMT translation projects, several strategies are employed:

  1. We regularly train our staff and external partners on how to use NMT and other AI tools responsibly. This training is part of our commitment to meeting ISO 27001 and 9001 standards.
  2. For our projects, we opt for “closed” proprietary or commercial NMT solutions and agree not to use NMT/AI systems for processing sensitive data.
  3. We work closely with our clients to clearly understand which data they consider confidential. We ensure all such data is handled with the utmost care, including signing the necessary agreements and adhering to specific workflows agreed upon with our clients.

These measures help us maintain the highest standards of data confidentiality and security in our translation projects.

Guidance for Businesses Using AI in Translation

Q: Finally, what advice would you give to businesses looking to leverage AI in their translation and localization efforts?

A: My advice to businesses wanting to use AI for translation and localization is pretty simple. Start by really understanding what you need. Not all translation tasks are the same, and AI might be great for some but not for others.

Next, remember that the human touch is crucial. The finest AI systems can still overlook crucial aspects. Hence, it’s wise to have experts review AI work, especially for critical documents or marketing content that must resonate on a personal level.

Thirdly, keep learning and adjusting. The world of AI is always changing, and what works today might be outdated tomorrow. Stay open to new ideas, keep testing how your translations perform, and be ready to switch things up to keep improving.

Lastly, talk to your customers and listen to their feedback. They’ll tell you if something doesn’t make sense or feels off. This feedback is gold for making your translations better and more effective. So, use AI as a tool to help you, but remember, it’s not a one-size-fits-all solution.

Combining technology with human expertise yields the best outcomes, whether through your internal team or external specialists.


Related Reading

Related Terms

Maria Webb
Technology Journalist
Maria Webb
Technology Journalist

Maria is a technology journalist with over five years of experience with a deep interest in AI and machine learning. She excels in data-driven journalism, making complex topics both accessible and engaging for her audience. Her work is prominently featured on Techopedia, Business2Community, and Eurostat, where she provides creative technical writing. She holds a Bachelor of Arts Honours in English and a Master of Science in Strategic Management and Digital Marketing from the University of Malta. Maria's background includes journalism for Newsbook.com.mt, covering a range of topics from local events to international tech trends.