AI language interpretation technology currently supports real-time interpretation for phone and video calls, with varying degrees of context awareness and voice clarity.

My motivation comes from personal experience as a first-generation American. Although my mother speaks English, her first language is Hindi, which has sometimes made communication with healthcare providers difficult. Clinicians often assumed she understood English perfectly, but this often was not the case, leading her to have difficulty complying with doctor's instructions. This experience motivated me to ensure that everyone who prefers a language other than English gets proper language services.

Initially, my mission in this space was to provide interpretation during clinical visits. But then I realized it was also needed during other phases of the patient journey. For instance, when my mother was given discharge instructions in English or when my dad received confusing medical bills. My team at GLOBO is working hard to ensure language services are available at every interaction during the patient journey.

Quality is of utmost importance to me for language interpretation. Right now, there are two main challenges in the AI interpretation space. First, ensuring that the interpreter interprets correctly and conveys the right empathy and emotion. The way information is delivered is just as impactful as whether that information is accurate. Second, the lack of industry-accepted quality standards. There are existing methodologies for assessing in-person interpretation, but these don’t adapt well for phone or video services. I believe that we first need to establish what good interpretation looks like and then measure AI against those standards.

I see three main concerns about AI interpretation coming from stakeholders. The first is hallucinations — mistakes where AI creates something nonsensical or off-topic. In medical interpretation, there’s no room for inaccurate interpretations that could put a patient’s life in danger.

Aside from hallucinations, healthcare providers are concerned about privacy and the appropriate tone for communication. When it comes to privacy, concerns center around where the vast amounts of data will be stored. Proper evaluation and testing of the tool are crucial in addressing these issues, as one can argue that using the tool is more private than having a random stranger listen in on the call.

In terms of communication and tone, AI could become more predictable than human interpreters because you know what to expect each time. On the other hand, there’s still uncertainty about AI’s ability to mimic human emotional nuance effectively and appropriately. This is crucial. When patients receive potentially life-altering information like a cancer diagnosis, they don’t want that to be conveyed in a robotic tone devoid of empathy.

Research shows that providing interpretation services can decrease patients' length of stay and lower readmission rates. Imagine being a patient who doesn't speak or understand English. If you can't communicate properly, the doctor might order unnecessary tests or misunderstand your condition, leading to longer stays and delayed treatment for your condition.

Interpretation services are also important for two-way communication between patients and providers on medication instructions, dietary needs, cultural preferences, and administrative conversations.

For example, a doctor may tell a patient to take their medication "once a day." If interpreted incorrectly using a simpler tool like Google Translate, this instruction could mean "eleven times a day" to a Spanish speaker. Without these services, patients might not take their medication correctly, which could lead to clinical repercussions. It can also be used to explain dietary restrictions in the hospital cafeteria, ensuring patients receive meals that adhere to their health and cultural needs.

AI interpretation can help patients schedule and show up to appointments. Research from Boston Children’s Hospital indicates that interpretation services reduce no-shows.

There’s also an opportunity to more efficiently match patients to either virtual or human interpretation services. We know firsthand that clinicians often wait a long time for human interpreters. This is especially true for languages that have fewer interpreters available at the ready. By using AI for basic interpretation tasks, live interpreters can be reserved for more complex conversations.

AI can support document translation, too. Many hospitals already translate critical documents such as discharge instructions; however, there is often a gap when it comes to translating other important materials, like bills. Not only can AI accurately translate the words on these documents, but it can also localize the language. In other words, it can adapt a document’s format and presentation to align with a patient's cultural preferences. For instance, some cultures prefer to see the total cost of a bill at the top of a page, and then the itemized breakdown, while other cultures prefer the opposite.

Investors are pouring a lot of money into companies developing their own interpretation products. Some companies are even creating ambient listening capabilities in multiple languages. This investment is substantial, and technology is continuously improving. I can’t predict how quickly AI interpretation will grow across the next 5-10 years, but I do see it becoming a critical part of the interpretation services landscape as capabilities expand.

Our goal at GLOBO is to evolve from being just a language interpretation company to a patient communications company. AI could be used to analyze the patient's words, cultural background, tone of voice, and visual expressions to provide feedback to the clinician. It could suggest that the patient didn’t understand, and that the conversation needs to be rephrased or simplified. AI could also act as a co-pilot for interpreters. I imagine interpreters working with AI in the background, which could prompt them to slow down, seek clarification, or reconsider the interpretation of a particular statement.

Hospitals tend to be more focused on cost rather than quality for interpretation services, but I believe the focus needs to shift more towards quality. Interpretation quality issues are usually raised in an ad hoc manner because hospitals lack a systematic way to assess quality. Quality shortcomings are noticed only if the clinician also speaks the patient’s native language, there’s a major miscommunication, or the patient’s family or interpretation vendor elevates an issue.

Unfortunately, quality standards for interpretation fall behind other industries and even other parts of the healthcare industry. For example, there are accepted methods to evaluate the accuracy of medical coding, including third-party audits, but the same is not true for interpretation. We need to move the needle for language interpretation services to have commonly accepted quality standards.

At GLOBO, we're working towards creating a system where hospitals can compare vendors based on both cost and quality in a quantifiable way. We believe it's crucial to ensure that whatever the doctor is saying is accurately interpreted since it ultimately impacts patient outcomes.

First, I believe AI will have a role in interpretation in the future. It might be difficult to predict exactly how it will unfold, but its role is promising. Second, when evaluating AI for interpretation, it's important to consider the various roles an interpreter can play. It's not just about delivering messages accurately but also speed and cost. Third, AI's potential extends beyond simply reducing costs. It can enhance language interpretation services in ways that improve overall quality, impact, and value for both clinicians and patients.

GLOBO Language Solutions ("GLOBO") is ranked in Nimdzi’s 2024 top 10 U.S. healthcare interpreting companies facilitating effective patient communication between healthcare providers and LEP patients. The company manages an independent global network of more than 8,000 linguists who speak 430+ languages and dialects. GLOBO supports leading healthcare organizations across the country through on-demand audio, video, on-site, and sign language interpreting; actionable insights; and translation of documents, emails, texts, and chats in a single AI-powered platform. GLOBO has been listed on the Inc. 500|5000 eight times and is a 2024 Vendors Division Semi-Finalist in Healthcare Innovation Magazine's annual Innovator Awards Program. Become a fan of GLOBO on LinkedIn.

Learn more about GLOBO.

This article is sponsored by GLOBO Language Solutions ("GLOBO"), an Advisory Board member organization. Representatives of GLOBO helped select the topics and issues addressed. Advisory Board experts wrote the article, maintained final editorial approval, and conducted the underlying research independently and objectively. Advisory Board does not endorse any company, organization, product or brand mentioned herein.

To learn more, view our editorial guidelines.