Artificial Intelligence: Test Highlights Human Interpreters' Irreplaceable Role

In a June test conducted with linguists Barry Slaughter Olsen and Walter Krochma, an artificial intelligence speech interpreter called Kudo was analyzed to determine its interpretation capabilities compared to those of a human. The focus of the test was not only on translating words but also on capturing the emotional nuances of the speech.

This was a follow-up to Olsen and Krochma’s famed 2019 video series in which they compared AI interpreters to human interpretation, concluding that AI still had a long distance to travel to catch up with human accuracy, cultural nuance, and conveyance of emotion.

Four years later, Olsen and Krochma once again aimed to maintain fairness and balance in assessing both the AI's performance and their own as interpreters. They ensured a level playing field by engaging in a preparation process similar to what a human interpreter would undertake before a task. This approach allowed for a comprehensive evaluation of the AI's performance and the comparison with human interpretation.

You can watch the 2023 video here.

The experience of creating the video not only provided educational insights but also challenged the professionals' prior assumptions and fears about AI technology. Krochma's reservations were particularly alleviated, and he gained a clearer understanding of how AI, when utilized appropriately and with human guidance, can be an invaluable tool that should not be feared.

At the same time, the experiment exposed the weaknesses and strengths of the technology. Though its accuracy had improved, the AI still trailed significantly in terms of nuance, cultural awareness, industry-specific terminology, technical terminology, and emotion.

Overall, the test highlighted the importance of human interpretation while acknowledging the potential value of AI as a complementary tool. It emphasized the need for ongoing learning, collaboration, and the utilization of AI in conjunction with human expertise for optimal results.

Six Reasons Human Interpretation Is Irreplaceable

AI has made remarkable advancements in various fields, transforming industries and revolutionizing how we work.

However, as the experiment indicates, there are fundamental reasons why AI is unlikely to fully replace human interpretation and translation when you also take into consideration the impact of machine translation and other large language models.

The nuances of human understanding, context, and empathy are crucial aspects that make human linguistics skills indispensable. Human interpretation and translation remain irreplaceable for reasons including:

1.       The Complexity of Language: Language is a rich and complex aspect of human communication that is constantly evolving. Understanding language involves more than just literal translation or pattern recognition. It requires grasping context, idiomatic expressions, cultural references, and subtle nuances. According to this technical report by OpenAI, current AI models still struggle with contextual understanding, often producing incorrect or nonsensical interpretations in complex linguistic scenarios.

2.   Emotional Intelligence and Empathy: Interpretation goes beyond words; it encompasses emotions, intentions, and underlying motivations. Human interpretation relies on emotional intelligence and empathy, enabling us to discern underlying feelings and provide appropriate responses. A survey by Accenture in 2016 found that 83% of consumers prefer human interaction over digital channels when dealing with emotional or complex situations, highlighting the value of empathy and understanding that AI cannot replicate it.

3.   Cultural and Societal Factors: Language is deeply intertwined with culture and society. Interpreting content requires a deep understanding of cultural nuances, historical references, and social context. And just like empathy, humans have the skills and experiences that surpass AI's ability to comprehend cultural intricacies.

4.   Creative and Critical Thinking: Interpretation often involves creative and critical thinking, particularly in fields such as literature, art, or legal analysis. AI excels at pattern recognition and data processing, but it struggles with the abstract and subjective aspects of interpretation. According to the World Economic Forum, critical thinking and creativity are among the top skills that will be in high demand in the future job market, underscoring the irreplaceable value of human linguistic skills.

5.   Ethical Considerations: Interpretation frequently involves ethical judgments and decision-making. Humans possess moral reasoning and ethical frameworks that shape their interpretations. AI lacks human ethical judgment and may produce biased or unethical interpretations. A study conducted by researchers at Princeton University and the University of Bath revealed that AI language models can amplify gender and racial biases present in the training data, emphasizing the need for human oversight in translation and interpretation.

6.   Adaptability and Flexibility: Language evolves, slang emerges, and new cultural references come forward over time. Human interpreters can adapt and update their understanding to keep up with these changes. AI, on the other hand, relies on pre-programmed algorithms and training data, limiting its ability to quickly adapt to dynamic linguistic and cultural shifts.

A Sum That Is Greater Than the Whole

While AI continues to advance rapidly, its limitations in translating and interpreting language are evident. Human involvement, with its emotional intelligence, cultural understanding, and ethical discernment, remains essential.

As we move forward, the integration of AI and human interpretation can lead to powerful collaborations, augmenting human capabilities and improving efficiency.

We must embrace the unique strengths of both AI and human interpretation to maximize our understanding. The focus of blending AI and human interpretation should be how best to relevantly engage both so, as Aristotle once said, the whole can be greater than the parts.

By recognizing the limitations of AI and embracing the unique strengths of human interpreters and translators, we can harness the potential of both to achieve new heights in understanding, communication, and innovation.

LanguageLine’s Approach

For more than 40 years, LanguageLine has excelled at melding human interpretation and translation with leading-edge technology. We view our position as the industry leader as a privilege and a mandate to be pioneers. As a company, we have a strong conviction in the power of language access to be transformative and strive to build new pathways to access.

LanguageLine’s team currently consists of over 35,000 linguists who provide interpretation and translation in more than 290 languages. Every day, in every way, we are accessing new technologies to help them do their jobs as well as they can be done.

The audiences our clients serve are growing more complex by the day. We don’t just meet their language access needs – we anticipate them by creating trailblazing solutions that simplify this complexity and turn challenges into opportunities.

Please contact us so that we can learn more about the opportunities that lie within your organization.