AI in medically trained translation tools
June 2026
Looking Back at Our Inspiration Session: AI Translation Tools in Healthcare – Opportunities, Challenges and Critical Success Factors
During our inspiration session on AI-powered translation tools in healthcare, we welcomed around thirty participants from hospitals, community health centres and general practice. The strong attendance and lively discussions confirmed that this is a highly relevant topic across the healthcare sector.
While some participants had already gained experience with AI translation tools, many were still seeking answers to questions regarding reliability, privacy, data protection, accuracy, potential AI hallucinations and erroneous interpretations.
Prof. Dr. Ellen Van Praet (Ghent University), dr. Pablo Decock and Fien De Wiest (HOGENT), researchers and coordinators of MATCHeN – Multilingual Access To Care & Health Expertise Network, provided a clear and balanced overview of the current state of AI-assisted translation in healthcare. Drawing on their research, they discussed the opportunities and limitations of the AI translation tools currently available to healthcare professionals.
One of the key conclusions was that AI translation tools specifically trained for medical use offer significant advantages over generic translation services such as Google Translate and general-purpose generative AI systems. Their focus on medical terminology and healthcare-specific contexts generally results in more accurate and reliable translations.
At the same time, the speakers emphasised that these specialised tools are not without limitations. The quality of AI-generated translations depends heavily on the context provided to the system and on the healthcare professional's ability to use the technology appropriately.
The session highlighted that the successful implementation of AI translation technology requires more than selecting the right application. Equally important is investing in the digital competencies of healthcare professionals, enabling them to understand both the capabilities and the limitations of these tools and to critically evaluate their output.
Which AI translation tool can currently be considered "the best"? The answer proved to be nuanced: there is no universal solution. The most appropriate tool depends on the specific healthcare setting, the languages most frequently encountered, and the functionalities that matter most to the organisation, such as speed, alternative translation suggestions or a back-translation feature that allows users to verify whether the intended message has been conveyed correctly.
The key takeaway from the evening was that AI translation tools can provide valuable support for multilingual communication in healthcare, but they cannot replace the clinical judgement and professional responsibility of healthcare providers. Actively verifying whether patients have correctly understood the information—for example, by asking them to explain the message in their own words—remains essential for delivering safe, high-quality and patient-centred care.
The session demonstrated that AI translation technology has the potential to make healthcare more accessible for increasingly diverse patient populations. Realising this potential, however, requires thoughtful implementation, adequate training, continuous critical reflection and an ongoing commitment to meaningful human interaction.
On 2 April 2026, THE INSTITUTE hosted an inspiration session on AI in healthcare. One message stood out clearly throughout the event: the maturity and impact of AI applications have increased significantly in a remarkably short period of time.
Through practical demonstrations by Bingli, Mindoo, Cavell, Squire, and Tiro.health, it became evident that AI is already delivering tangible value in everyday clinical practice. Looking ahead, AI is expected to evolve even further over the coming months - from providing operational support to acting as a genuine assistant for healthcare professionals.
One of the most powerful insights came directly from physicians themselves: AI has the potential to make consultations enjoyable again. Not because technology replaces the clinician, but because it can substantially reduce the administrative burden. This allows healthcare professionals to focus once again on what matters most: spending more time and giving more attention to their patients.
Applications in medical history taking (anamnesis) and triage were identified as having particularly high potential. At the same time, it became clear that building a business case is currently easier in the United States than in Belgium. For example, expanding triage capacity in the U.S. often translates directly into additional revenue. Nevertheless, the overall direction is clear: AI is rapidly evolving into a personal assistant for healthcare professionals.
Adoption is also accelerating at an impressive pace. In the United States, healthcare has become the sector with the highest rate of AI adoption. This momentum is largely driven by use cases with an obvious return on investment, such as administrative support and medical billing.
AI hallucinations naturally remain an important point of attention. However, the session also demonstrated that their impact can be significantly reduced through validation processes, feedback loops, and the use of well-designed templates.
Another important challenge emerged during the discussions. Physicians and other healthcare professionals expect AI systems to provide explainable reasoning, whereas many AI models operate as "black boxes." This lack of explainability can make it difficult for clinicians to fully trust AI-generated recommendations.
Today, the main barriers to large-scale adoption are no longer the technology itself, but rather the surrounding ecosystem: affordability for hospitals, complex integration with existing Electronic Health Record (EHR) systems, competition with other IT priorities, fragmentation across healthcare disciplines, and the need for a compelling business case for physicians, hospitals, and patients alike.
For THE INSTITUTE, one key takeaway from this session is that further integration of AI into healthcare will require greater awareness, broader engagement of the medical community, closer collaboration between physicians and hospitals, and a clearer narrative about AI's impact on quality of care and value creation.
Perhaps we should also rethink the way we talk about AI. Rather than focusing primarily on efficiency gains, we should emphasise what AI ultimately enables: more time for meaningful patient interaction.
A sincere thank you to all speakers and participants for their openness, valuable insights, and inspiring discussions. The future of AI in healthcare will not be shaped by technology alone, but by how we choose to implement it together to make healthcare more human, higher in quality, and more sustainable for those delivering it.
The healthcare assistant that thinks along with the healthcare provider: asks for and records basic information before the patient arrives.
Smart medical, AI-driven medical history and care navigation for patients and healthcare providers.
Smart medical documentation that converts time-consuming administrative work into valuable, structured data.
Listens in on the consultation and summarises the conversation using AI. It also generates complete consultation letters, hospitalisation letters, referral letters, certificates, etc. based on a concise quote.
Squire supports doctors and nurses throughout the consultation process: it saves time by speaking to the patient in advance, generates medical documentation in the healthcare provider’s preferred style and language by listening in, and allows findings to be dictated to quickly complete documents or EPR fields.