Over 25 million people in the United States have limited English proficiency, creating a critical need for reliable medical interpretation. Traditional in-person interpretation can be logistically challenging and costly, especially in urgent or remote settings. Consequently, remote and AI-enabled language interpretation services have emerged as vital tools to bridge communication gaps in clinical environments.

What Are Remote and AI-Enabled Language Interpretation Services?

Remote and AI-enabled language interpretation services facilitate real-time language interpretation in clinical settings through phone- or video-based, on-demand human interpreters or AI-driven machine translation technology. Given the stringent requirements for HIPAA compliance, enterprise security, and the need to accurately interpret medical terminology, consumer-oriented translation services are often unsuitable for healthcare applications. 

Products in this category are designed to meet these specific needs, ensuring secure and precise communication between healthcare providers and patients. These solutions are distinct from translation services, which are asynchronous and text-based, though some vendors offer both.

Historically, medical interpretation relied heavily on in-person interpreters, which, while effective, present challenges in terms of availability, especially in diverse linguistic regions or during emergencies. Telephonic interpretation was popularized during the ‘80s and ‘90s, but skyrocketed during the COVID pandemic as organizations aimed to reduce in-person interactions. During and since that time, advancements in video conferencing—which gives interpreters visual and contextual information in addition to verbal input—and artificial intelligence have allowed these solutions to become more prevalent.

Vendor Landscape

Modern medical interpretation services operate through two primary modalities:

Remote Human Interpretation: Utilizing video remote interpreting (VRI) platforms, these services connect patients and healthcare providers with professional interpreters via secure video links. This approach maintains the nuances of face-to-face communication, ensuring that non-verbal cues are preserved. The majority of the solutions within our category are remote, human interpretation services. 

While these solutions generally work well, proponents of AI interpretation argue that human translators may paraphrase or unconsciously introduce bias into their communications, whereas AI can be more accurate and allow the provider and patient to speak more naturally without pauses for interpretation. Additionally, connecting to a human can be time-consuming and cumbersome when providers have limited face-time with patients. 

AI-Powered Interpretation: AI-driven services such as Jaide Health, Mabel AI, No Barrier, and Sully.ai provide instantaneous translation without human intermediaries. 

But while these systems offer speed and scalability, their accuracy in medical contexts is a subject of ongoing concern. Practically speaking (assuming patient and provider have no shared language), AI interpretation is an entirely autonomous application of AI without any human oversight. Some express concern that misinterpretations in critical medical information by AI systems could pose significant risks, leading to potential misdiagnoses or inappropriate treatments and introducing complex liability considerations. There is not yet a legal precedence for liability here, making providers hesitant. 

Additional differentiators: Beyond this primary distinction, additional differentiators across both remote human and AI services include document and message translation services, EHR integration, and integration with telehealth and video-conferencing platforms.

Future Directions

While factors like usability, wait time, and integrations may somewhat influence vendor selection, in order for AI medical translation and interpretation to really take hold, two key advancements need to occur:

• Technological: AI vendors must establish that their translation services are quantifiably more accurate than human translation. Usability or convenience simply can’t overcome the fact that outputs could be inaccurate, particularly when providers aren’t able to verify accuracy at the point of care. As empathetic language continues to grow as a differentiator among LLMs, the ability to integrate cultural nuances and empathy will likely also be valued but only as a secondary consideration.
  

• Legal Precedence: Governance and liability considerations are obviously major issues that healthcare institutions are grappling with more broadly. These will undoubtedly come to a head over the coming years (if not months), but until then, we expect that most clinicians will remain slow to adopt.
  

In the meantime hybrid models that combine AI efficiency with human oversight may emerge as the optimal approach. For example, these solutions may take an initial pass on document translation or transcribing clinical notes from patient-provider encounters, allowing interpreters to simply refine.