Over 500 Medical Affairs professionals from around the world arrived in Zurich, Switzerland, for the annual Medical Affairs Professional Society (MAPS) EMEA meeting on 17–19 May 2026. One of the strongest recurring themes was the increasing complexity of Medical Affairs’ role in communicating its science and how the environment it operates in is rapidly evolving.  

The information environment is changing faster than Medical Affairs can respond with current operating models

HCPs, patients and payers are currently navigating an overload of health information, as well as combatting ‘MDM’ - misinformation, disinformation and malinformation. WHO describes this as an infodemic - an overabundance of information, including false or misleading information, that creates confusion, harmful risk-taking and mistrust in health authorities, and undermines public health messaging. Social media, peer communities and digital opinion leaders (DOLs) are shaping what people notice, trust and search for, often in real time – often before they encounter formal medical education or original content. This creates a risk that misinformation is not only consumed, but also repeated and reinforced within professional conversations.

AI adds another layer of complexity. These platforms increasingly act as informal health information intermediaries, summarising and re-presenting what is most visible online. However, they may miss restricted (for example behind paywalls, journal access, language, geography or local information pathways), emerging or nuanced scientific context, making incomplete information appear more certain than it is.

For Medical Affairs, scientific accuracy remains essential, but it is no longer enough. Trusted science now needs to be accurate, but also timely, visible, findable and structured for interpretation across this evolving information ecosystem.

Current governance and review processes focus on scientific accuracy, balance, compliance and evidence-based exchange. However, this rigour often means content and health guidance is outpaced by the rapidly evolving online conversations. Current approaches to guiding scientific conversations are limited if they do not account for wider information ecosystems.

Medical Affairs needs to be responsible for designing a trusted architecture around the science of its medicines – the answer is not more content but a designed, agile system

This designed system faces the challenge of the volume, fragmentation and uneven visibility of scientific information across journals, congresses, guidelines, social platforms, peer networks, patient communities and AI-generated summaries. In this environment, even accurate science can fail to influence decisions if it is inaccessible, poorly structured, hard to interpret, or not surfaced by the platforms and tools people use to find answers.  

Our recent podcast, ‘One size does not fit all – insights for meaningful DOL partnerships to counter global misinformation’ touches on this challenge. How we respond to misinformation still matters substantially – helping people think critically, slowing the spread of misinformation and correcting false claims – but they are no longer enough on their own.  

Medical Affairs, therefore, needs to build a ‘trusted architecture’ for scientific exchange, using a connected system of credible evidence, clear scientific narratives, expert interpretation and structured content. This includes understanding where audiences are looking for information, which sources they trust, which voices shape interpretation, and where evidence is distorted, diluted or simply absent across the information journey. Delivering this trusted architecture requires a more agile operating model that supports faster content release and the rapid resolution of evidence gaps.  

AI is becoming a new audience for Medical Affairs

Patients, HCPs and payers increasingly use AI tools to ask questions, summarise evidence, understand options and pressure-test decisions. However, AI chatbots do not automatically ‘see’ the full landscape of credible scientific context. An audit in BMJ Open¹ tested five major public-facing AI chatbots (including ChatGPT, Gemini, Deepseek, Meta AI and Grok) across 250 health and medical queries using an adversarial framework designed to mirror how real people ask questions. Around half of responses (49.6%) were classed as problematic and 1 in 5 (19.6%) as highly problematic. The chatbot refused to answer only two (0.8%) of 250 prompts. The same audit found that the quality of references produced was poor, with an average reference completeness score of 40%. Chatbot hallucinations or fabricated citations meant that no chatbot produced a fully accurate reference list, although these responses were often delivered with confidence and certainty.  

The structural issue is that general-purpose chatbots do not reason or weigh evidence as clinicians do. They predict likely word sequences from accessible training data and retrievable content, which can skew toward open-access material, Q&A forums and social media. Publicly available LLMs do not account for missing, specialist or emerging evidence, nor whether the information should be classified as ‘MDM’. This reinforces the need to view current AI tools not as a decision-maker in healthcare, but as a new intermediary and an increasingly influential filter between evidence and the people using it.

A mindset shift is needed:

What this means for Medical Affairs  

Medical Affairs has always been a guardian of scientific accuracy for its medicines. In the next era, it must also become a guardian of rapid scientific visibility for them. This requires a new approach that integrates medical strategy, scientific storytelling, influencer and discourse mapping, social intelligence, AI-readiness and fast content governance.  

To succeed Medical Affairs will have to build relationships with previously unengaged specialists, such as data analysts and strategists, and technical engineers.

It will require creating tools that help Medical Affairs understand where misinformation or misinterpretation is emerging; which voices and platforms are shaping belief; what credible science is visible or invisible to AI chatbots; and where content needs to be restructured, clarified or activated. If Medical Affairs does not optimise credible science for AI, less credible voices may fill the gap.  

At Langland Medical we believe the future of scientific exchange will be won by teams that make trusted science not only accurate, but visible, findable and understood – by people and by the AI systems increasingly shaping what people see. If you’d like to hear more about how we have evolved to help clients navigate this new information landscape, let's continue the conversation.

‍

Reference

1. Tiller NB, et al. BMJ Open. 2026;16:e112695.

‍