{"id":325,"date":"2026-06-02T15:15:52","date_gmt":"2026-06-02T15:15:52","guid":{"rendered":"https:\/\/facultyhub.elsevier.com\/en\/?p=325"},"modified":"2026-06-19T14:05:24","modified_gmt":"2026-06-19T14:05:24","slug":"raman-kaur-on-building-trustworthy-ai-for-medical-education","status":"publish","type":"post","link":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education","title":{"rendered":"Raman Kaur on Building Trustworthy AI for Medical Education"},"content":{"rendered":"\n

Elsevier Medical Education presents an interview with Raman Kaur, Vice President in Education Technology at Elsevier Health, focused on how trustworthy, evidence-based AI can support medical education, exploring accuracy and bias, hallucination safeguards, human oversight, and data protection.<\/strong><\/p>\n\n\n\n

For those who are less familiar with the topic, could you briefly explain how AI integrates reliable data for medical education? <\/h2>\n\n\n\n

AI in medical education works best when it’s grounded in trusted medical content and not just open internet data. I think that’s already been studied and proven \u2013 that connecting our education tools with trusted content creates that epistemic trust with our educators and with our students, and that’s a key competitive advantage for any tool. So, in practice, the system combines large language models, known more popularly as generative AI, with curated educational resources such as peer-reviewed textbooks, clinical references, anatomy content and assessment materials. Along with that, we like to incorporate educator-reviewed learning experiences as well, because what’s better than the real life experiences of an educator for making the AI aspect of the medical education more grounded and trustworthy? Similar to Osmosis AI from Elsevier, user AI-driven experience helps students reinforce concepts through videos, flashcards, question banks, adaptive learning support, and a conversational AI that sits on top of all of these capabilities. The value comes with combining the AI capability with the trusted educational content and learner engagement data to make it a very powerful, tailored, and curated experience.<\/p>\n\n\n\n

If I could give one more example via our products, such as ClinicalKey Student, the learners can access evidence-based medical references as well as textbooks, videos, and clinical answers in one ecosystem by having an all-in-one experience instead of having to click through different tabs, links, and content delivery formats. AI can help surface the most relevant information faster and support more personalized learning pathways. So, this is generally how we’re approaching AI integration with medical education, with that epistemic trust in the middle of it.<\/p>\n\n\n\n

How is the accuracy and potential bias of this content evaluated, and what tools are used for this purpose? <\/h2>\n\n\n\n

I think with evaluation being one of our most important parts of the healthcare AI education incorporation, we feel very comfortable adding AI to our tools. So, what does evaluation mean? Evaluation is essentially our human-in-the-loop process. So, as we introduce more AI capabilities to our medical education platforms, there is a constant human-in-the-loop AI evaluation to make sure that we identify and catch any type of inaccuracy or potential bias from happening before it lands in the hands of the students. Evaluation is actually one of the most important parts of making sure that an AI tool is grounded and is providing you with answers that predominantly look at the information constructively and are free of any type of hallucinations. For example, if a student is using an AI support experience inside ClinicalKey Student to understand a cardiac condition, the system needs to provide information that’s aligned with trusted references as well as medical standards. So, organizations will evaluate things like factual accuracy, clinical safety, citation quality, and bias across patients’ populations. We look for all these types of challenges that could happen when AI is delivering an answer to make sure that our tools remain grounded and trusted and deliver ethically sound and true education-aligned answers. <\/p>\n\n\n\n

There is also a strong human review component in it. We will just not be releasing our tools \u2013 we will have that human review take place. We want to see a certain type of review score. We want to see a certain type of accuracy and bias-free score before it will enter the ecosystem of medical knowledge. Also to keep in mind, the idea of evaluation is continuous. So, what I want people to take away is that once the evaluation happens for a version of an AI tool in medical education, we don’t stop there. This evaluation happens on an ad hoc basis and on a consistent basis because medical education is a rapidly evolving construct. So, AI systems and healthcare education cannot just set and forget. We do evaluations consistently.<\/p>\n\n\n\n

\n

\"One of the biggest risks in healthcare AI is not that the system sounds uncertain \u2013 it’s actually when the system is sounding extremely confident while being wrong. This is why we have a special evaluation framework for hallucinations.\"<\/h2>\n<\/section>\n\n\n\n

Also, for non-experts: what are \u201challucinations\u201d in AI, and how can they hinder the effective use of AI in medical education? <\/h2>\n\n\n\n

In respect to AI, \"hallucination\" occurs when AI systems generate information that sounds convincing but it’s not actually correct. It’s actually more fabricated, put together pieces of information that, again, makes you think that it sounds about right, but it isn’t quite there and if you were to quickly glance at it, you might even miss that it’s a hallucinated piece of artifact and not an accurate piece of information. So, in medical education, it could look like inventing a clinical citation for medical research or other disciplines of research. This is a very popular challenge. Also, DOI IDs are always being made up in recent times. It’s getting better but that would be one giveaway. Or giving an inaccurate explanation of a disease process because you connected two similarly sounding symptoms into one. Also, suggesting outdated treatment guidance. So, say our treatment guidance was updated last year, it may pull from a previous treatment guidance. Or even mislabeling anatomical structures because remember, human brain processes process more context than AI can. So, it might label an epicondyle next to a bicep, for example. This is where a hallucination can happen.<\/p>\n\n\n\n

Then we have to keep in mind that catching these hallucinations through the evaluation process that I mentioned before is highly critical because individuals who are consuming our healthcare AI education materials are not experts yet. So, we wouldn’t want them to ground their learning in inaccurate systems. That’s why trustworthy healthcare AI system rely heavily on the things I mentioned before, like grounded medical content, citation and traceability, human oversight, consistent evaluations, and safety monitoring. <\/p>\n\n\n\n

Safety monitoring is another one that is probably unique to healthcare education tools versus, say, an AI tool for sales. The safety piece of it is really important because you want to make sure that the information that’s being gained out of our tool is relevant, is the latest, and is the right method of potentially building knowledge of a future treatment or a future surgery intervention. One of the biggest risks in healthcare AI is not that the system sounds uncertain \u2013 it’s actually when the system is sounding extremely confident while being wrong. This is why we have a special evaluation framework for hallucinations.<\/p>\n\n\n\n

What are the key characteristics or requirements of a trustworthy AI architecture – both in terms of content and implementation – within medical or healthcare education? <\/h2>\n\n\n\n

This is basically what we live and breathe every day. How do we make AI solutions trustworthy in healthcare education and build a sustainable architecture of maintaining that level of trustworthiness? To ground us, what does a trustworthy AI system look like? In respect of medical education or any healthcare education, it starts with strong governance around the content, as well as incorporating the trusted sources where AI is referencing the content from. For example, ClinicalKey Student brings together authoritative medical references and a structured learning resource environment. It makes a very powerful combination. For Osmosis, it is reinforcement and a personalized learning experience through tailored video content. Complete Anatomy, of course, is one of our favorites. It provides a highly visual interactive anatomy education and allows you to essentially identify structures that you otherwise would not have access to, even if you were to be in a high-fidelity lab setting.<\/p>\n\n\n\n

So, an AI layer should really enhance those trusted experiences and not bypass them, right? There are several characteristics that we can identify in tools that we would call trustworthy in medical education. Again, and I know I keep saying this, but the evidence-based content is so critical. Also, transparent sourcing of citations so that you know that your learning is coming from a traceable source that is grounded in reality. The human oversight part is essential. There should never be an assumption that any AI tool in education can just replace a faculty member. That’s just not a thing. It also, from an architectural setting and a sustainable architecture standpoint, has a lot of privacy and security protections. Why that\u2019s important is that there will be times that our content is something that we want to make discoverable or it becomes discoverable without our decision. So, that privacy and security protection of the AI ecosystem is very critical because our content is what’s powering our AI tools, and we wouldn’t want that content used in the wrong setting because context means everything. Ongoing monitoring and evaluation continues to be a very essential part of the sustainable architecture of a trustworthy AI tool for medical education. And then lastly, reliable system performance. How many times has an AI tool told you, \u201cI’m out of capacity,\u201d \u201cI don’t know the answer to that,\u201d or \u201cI’m sorry, I can’t help you\u201d? Those experiences degrade user perception of the tool so we have to make sure that our architecture supports sufficient token capacity and processing speed to not let our users down.<\/p>\n\n\n\n

So, to summarize in two main points: firstly, trustworthy AI should be supporting critical thinking, not replacing it. So, challenging the user to think on their own with the right kind of information and context. And secondly, the best educational AI system guides the learners towards a deep understanding instead of just simply generating and spitting answers at them. So, if you wanted to do a quick evaluation, these two things will tell you that this is a trustworthy and reliable AI tool for learning in medical settings.<\/p>\n\n\n\n

What challenges does data protection currently pose for the use of AI in the context of medical education? <\/h2>\n\n\n\n

So, remember when we talked about the hallucination piece? Well, what’s happening today is that certain content leakage issues, which ties into the data protection, allows some of the open-source AI tools to piece information together like Frankenstein’s bride, and that obviously promotes more hallucinations and more biases. That’s why our focus is heavily on \u2013 obviously innovation \u2013 but also protecting our data and creating the right data pipelines that connect with our AI tools. Essentially, it’s as simple as \"what is your skillset and what is the AI tool you’re using?\" So, if I’m a medical student and I’m using Osmosis AI, I can be confident that I’m getting the right information because I know there’s trusted content and clear, defined pipelines to that content that sit behind my Osmosis AI experience. However, if I go to an open-source AI tool that is not meant for medical education, I cannot guarantee that I will be getting the most accurate, grounded, ethically sound, bias-free and hallucination-free answer. So, the skillset mapping to the right use of the right AI tool is extremely important, and that’s where the data protection comes into play.<\/p>\n\n\n\n

Data protection becomes a very major consideration for us because healthcare AI tools obviously involve sensitive patient information, and medical education involves delivering the content in the right setting because these will be people who are future clinicians. Institutions that are using AI-enabled learning platforms with high confidence want to make sure that they have student privacy put in place and that they have regulatory compliance. You may have different guidelines being practiced in, say, North America versus a European region. Even within the countries, different regions may follow different guidelines. So, that trustworthiness and the data protection and data pipelining is very important. Additionally, third-party model use is a really big thing. Of course, like any other fast-moving, innovative organizations, we have a lot of third-party vendors that we use to help us get to market faster. However, the segmentation of data between them and the creation of a secure gateway that only allows capabilities to be shared and not data is a very important aspect of our applications and learning platforms. Retention policies are also really important. We retain some level of prompt information to make sure if there were to be a future report of a hallucination or a bias, that we can go back and look at what the interaction was like. And then lastly, security controls. We would only want licensed users who are supposed to be accessing the tool to have access to our platform. It really shouldn’t be a random person launching our learning website and trying to look for answers. I think that’s not the right use of our information, which is why data protection comes into play.<\/p>\n\n\n\n

So, to tie everything together, there is obviously a growing need to ensure that sensitive institutional information or clinical information is well protected and it’s not unintentionally exposed through AI workflows to the open internet. I also think that the future will involve stronger governance models. I know things are evolving very rapidly today but governance, unfortunately, always kind of follows at the heels of it. We see this happening in real time now where customers are demanding governance patterns to ensure that their data is protected through our AI tools, and we have the receipts to be able to showcase that. And lastly, how do we preserve privacy through the AI architecture and a clear transparency of our standards, which is a demand across industry today.<\/p>\n\n\n\n

In your opinion, what are the main trends or directions that will shape the future of AI in healthcare education? <\/h2>\n\n\n\n

There are so many things but if I had to summarize, I would say there are three trends that we’re seeing. Personalization is a big one \u2013 the right skill, the right user, the right AI tool. The personalization piece is becoming very important because a generic AI tool will throw a lot of information at you and I don’t think that we have time to weed through the essays and essays of blurbs that come out of a generic tool. So, that personalized experience and increased tailoring of learning experiences to the individual is extremely important. That’s becoming more apparent with our individual current student users, faculty users, as well as our potential clients.<\/p>\n\n\n\n

Second is multimodal learning. We all know that the cost of AI is still fairly expensive. It could be cost prohibitive for certain smaller institutions to consume AI-driven learning experiences. So, we’re moving beyond just text into more immersive and interactive experiences that then further creates the token cost arbitrage. Essentially, what we’re looking to do is go a level deeper in the refinement of how we use our language models and use a certain language model for what it is good at. You’ve probably seen in some of the open-source generative AI tools, you can choose from \u201cinstant answer\u201d, \u201cthinking answer\u201d, \u201cresearch answer\u201d. At Elsevier, we’re kind of taking a similar approach. So, if you’re just asking a quick question and we can use our proprietary system to go retrieve an answer, that’s a very small, low-cost model. But if you’re asking it to do a literature review on a certain concept, that’s a \u201cthinking model\u201d. We are also looking at opportunities to create that very curated pathway of multi-model architecture.<\/p>\n\n\n\n

Then lastly, workflow integration. Nobody wants to leave their workflow, click out of what they’re doing, and go click into another platform. How do we bring the trusted AI learning experiences via our tools into a student’s workflow so that it is part of their day-to-day living and breathing experience. That’s another emerging trend that we’re seeing so that it is a natural thing for them to just turn towards a certain experience instead of having to rely on their traditional experiences.<\/p>\n\n\n\n

I think these three are extremely important. If I had to add a bonus one, there\u2019s another trend that we’re seeing in healthcare education which is the demand for AI literacy within healthcare professionals. So, imagine training our students to be this highly proficient, AI-fluent future workforce of tomorrow, but then when they enter the workforce, their mentors in the field are not AI fluent. That’s such a distant experience. So, we’re also seeing a motivation among healthcare professionals to get more understanding of AI tools so that, as our future clinicians who are students today enter the workforce, they can continue to enjoy those efficiencies and that improved user experience from their future systems, like the point of care system.<\/p>\n","protected":false},"excerpt":{"rendered":"

Elsevier Medical Education presents an interview with Raman Kaur, Vice President in Education Technology at Elsevier Health, focused on how trustworthy, evidence-based AI can support medical education, exploring accuracy and bias, hallucination safeguards, human oversight, and data protection. <\/p>\n","protected":false},"author":289,"featured_media":606,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"event_date":"","event_end_date":"","event_location":"","event_cta_text":"","event_form_url":"","event_bigmarker_id":"","testimonial_name":"","testimonial_role":"","testimonial_institution":"","testimonial_country":"","_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[2,3],"tags":[],"class_list":["post-325","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog","category-medical"],"yoast_head":"\nRaman Kaur on Building Trustworthy AI for Medical Education - Faculty Hub EN<\/title>\n<meta name=\"description\" content=\"Raman Kaur explains how trustworthy AI can be integrated into medical education, addressing accuracy, bias, hallucinations, and privacy.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Raman Kaur on Building Trustworthy AI for Medical Education - Faculty Hub EN\" \/>\n<meta property=\"og:description\" content=\"Raman Kaur explains how trustworthy AI can be integrated into medical education, addressing accuracy, bias, hallucinations, and privacy.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education\" \/>\n<meta property=\"og:site_name\" content=\"Faculty Hub EN\" \/>\n<meta property=\"article:published_time\" content=\"2026-06-02T15:15:52+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-06-19T14:05:24+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1080\" \/>\n\t<meta property=\"og:image:height\" content=\"540\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"abarnewell\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"abarnewell\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"13 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education\",\"url\":\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education\",\"name\":\"Raman Kaur on Building Trustworthy AI for Medical Education - Faculty Hub EN\",\"isPartOf\":{\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#primaryimage\"},\"image\":{\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#primaryimage\"},\"thumbnailUrl\":\"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png\",\"datePublished\":\"2026-06-02T15:15:52+00:00\",\"dateModified\":\"2026-06-19T14:05:24+00:00\",\"author\":{\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/#\/schema\/person\/53ac6bd74f83cb02a31860a911e8d45c\"},\"description\":\"Raman Kaur explains how trustworthy AI can be integrated into medical education, addressing accuracy, bias, hallucinations, and privacy.\",\"breadcrumb\":{\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#primaryimage\",\"url\":\"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png\",\"contentUrl\":\"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png\",\"width\":1080,\"height\":540},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/facultyhub.elsevier.com\/en\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Raman Kaur on Building Trustworthy AI for Medical Education\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/#website\",\"url\":\"https:\/\/facultyhub.elsevier.com\/en\/\",\"name\":\"Faculty Hub EN\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/facultyhub.elsevier.com\/en\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/#\/schema\/person\/53ac6bd74f83cb02a31860a911e8d45c\",\"name\":\"abarnewell\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/facultyhub.elsevier.com\/en\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/493457a6063c62b90c84fb11a80e2a7bf870ae421f23f260f504d0fc39915e68?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/493457a6063c62b90c84fb11a80e2a7bf870ae421f23f260f504d0fc39915e68?s=96&d=mm&r=g\",\"caption\":\"abarnewell\"},\"url\":\"https:\/\/facultyhub.elsevier.com\/en\/author\/abarnewell\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Raman Kaur on Building Trustworthy AI for Medical Education - Faculty Hub EN","description":"Raman Kaur explains how trustworthy AI can be integrated into medical education, addressing accuracy, bias, hallucinations, and privacy.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education","og_locale":"en_US","og_type":"article","og_title":"Raman Kaur on Building Trustworthy AI for Medical Education - Faculty Hub EN","og_description":"Raman Kaur explains how trustworthy AI can be integrated into medical education, addressing accuracy, bias, hallucinations, and privacy.","og_url":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education","og_site_name":"Faculty Hub EN","article_published_time":"2026-06-02T15:15:52+00:00","article_modified_time":"2026-06-19T14:05:24+00:00","og_image":[{"width":1080,"height":540,"url":"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png","type":"image\/png"}],"author":"abarnewell","twitter_card":"summary_large_image","twitter_misc":{"Written by":"abarnewell","Est. reading time":"13 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education","url":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education","name":"Raman Kaur on Building Trustworthy AI for Medical Education - Faculty Hub EN","isPartOf":{"@id":"https:\/\/facultyhub.elsevier.com\/en\/#website"},"primaryImageOfPage":{"@id":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#primaryimage"},"image":{"@id":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#primaryimage"},"thumbnailUrl":"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png","datePublished":"2026-06-02T15:15:52+00:00","dateModified":"2026-06-19T14:05:24+00:00","author":{"@id":"https:\/\/facultyhub.elsevier.com\/en\/#\/schema\/person\/53ac6bd74f83cb02a31860a911e8d45c"},"description":"Raman Kaur explains how trustworthy AI can be integrated into medical education, addressing accuracy, bias, hallucinations, and privacy.","breadcrumb":{"@id":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#primaryimage","url":"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png","contentUrl":"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png","width":1080,"height":540},{"@type":"BreadcrumbList","@id":"https:\/\/facultyhub.elsevier.com\/en\/raman-kaur-on-building-trustworthy-ai-for-medical-education#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/facultyhub.elsevier.com\/en\/"},{"@type":"ListItem","position":2,"name":"Raman Kaur on Building Trustworthy AI for Medical Education"}]},{"@type":"WebSite","@id":"https:\/\/facultyhub.elsevier.com\/en\/#website","url":"https:\/\/facultyhub.elsevier.com\/en\/","name":"Faculty Hub EN","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/facultyhub.elsevier.com\/en\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/facultyhub.elsevier.com\/en\/#\/schema\/person\/53ac6bd74f83cb02a31860a911e8d45c","name":"abarnewell","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/facultyhub.elsevier.com\/en\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/493457a6063c62b90c84fb11a80e2a7bf870ae421f23f260f504d0fc39915e68?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/493457a6063c62b90c84fb11a80e2a7bf870ae421f23f260f504d0fc39915e68?s=96&d=mm&r=g","caption":"abarnewell"},"url":"https:\/\/facultyhub.elsevier.com\/en\/author\/abarnewell"}]}},"jetpack_featured_media_url":"https:\/\/facultyhub.elsevier.com\/en\/wp-content\/uploads\/sites\/29\/2026\/06\/26_2269_AB_Headshot_Blogs-03.png","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/posts\/325","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/users\/289"}],"replies":[{"embeddable":true,"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/comments?post=325"}],"version-history":[{"count":6,"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/posts\/325\/revisions"}],"predecessor-version":[{"id":935,"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/posts\/325\/revisions\/935"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/media\/606"}],"wp:attachment":[{"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/media?parent=325"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/categories?post=325"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/facultyhub.elsevier.com\/en\/wp-json\/wp\/v2\/tags?post=325"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}