Image Source: Getty Images
The COVID-19 pandemic highlighted how biosecurity crises depend heavily on scientific communication and trust. Public health advice, scientific research, political rhetoric, and digital media became interdependent, thus making information governance an important aspect of pandemic readiness. Decisions had to be made based on evolving evidence, while at the same time, information on scientific results, and at times misinformation about health, was spreading through social media.
With the advent of artificial intelligence (AI) and its many offshoots into different areas of science, biosecurity is no exception. AI and its impact on biological agents are often discussed; there is also the increasing use of AI in open-source early warning systems that utilise different platforms to inform the Early Warning Systems (EWS) output. The question then arises: how do we differentiate between scientific communication, misinformation, disinformation, and dissent when platforms increasingly used for crisis management are also being used to detect relevant signals?
As such, the World Health Organization (WHO) characterised this as an “infodemic”, referring to the coexistence of true and false information during health emergencies, making it difficult to discern the difference.
COVID-19 and the Information Challenge
COVID-19 pandemic highlighted the challenges associated with scientific communication in crisis circumstances. Initial doubts about severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) led to changes in recommendations concerning airborne spread, mask use, vaccination strategies, and public health restrictions. Although such changes were justified by new data, they were often perceived as inconsistent across multiple reporting platforms and word-of-mouth advice. The coexistence of reliable scientific knowledge and untruthful claims on both official and unofficial digital channels compounded the problem. Claims about vaccines, medical treatments, and even the existence of the coronavirus itself reached large audiences, while discussions about quarantine, school closures, and other aspects of daily life became increasingly political.
Scientific disagreement does not necessarily amount to misinformation. Traditionally, scientific advancement entails formulating hypotheses, conducting peer review, replicating results, and comparing competing explanations. In the case of outbreaks, different scientific understandings may coexist in the early stages of knowledge accumulation. The discussion over the role of airborne transmission illustrates this. of this. At the start of the pandemic, official advice concentrated on the danger of transmission through respiratory droplets and contaminated surfaces. At the same time, aerosol scientists emphasised the greater role of airborne transmission. The accumulation of new evidence led to changes in the guidelines, reflecting the evolving scientific understanding.
However, the information environment has become much more complicated due to the growing popularity of Open-Source Intelligence (OSINT) and Artificial Intelligence (AI)-based approaches to biosecurity surveillance. Modern AI-assisted systems analyse vast amounts of data provided by social media sites, news sources, blogs, web forums, and other publicly available information to identify signs of emerging outbreaks before they are detected through traditional surveillance mechanisms. While this increases situational awareness and can help detect outbreaks faster, this approach opens the way for an influx of misinformation, disinformation and even influence operations.
As the application of AI technologies becomes more widespread in biosecurity, the quality of open-source information will be as crucial as improvements in AI’s technological performance.
During the pandemic, various claims regarding the origin of the disease, vaccines, therapies, and other health policies spread rapidly online, outpacing official announcements. If such information is processed by AI-assisted systems without adequate verification, it may lead to erroneous risk assessments, false alarms, the masking of real outbreak signals, and the spread of fake news. As the application of AI technologies becomes more widespread in biosecurity, the quality of open-source information will be as crucial as improvements in AI’s technological performance.
Communication challenges during a pandemic persist with outbreaks such as mpox and highly pathogenic avian influenza (H5N1). The mpox outbreak in 2022 illustrated the inseparability of information-sharing and issues of stigma and trust. Given that the disease emerged within specific sexual networks, public health communication on the spread of the virus had to avoid deepening stigma or discouraging testing and treatment. This highlights another challenge in biosecurity communication: even technically accurate information can harm public trust if presented without due regard for context.
Similarly, authorities monitoring H5N1 have had to balance preparedness with proportionate communication. Although human infections have occurred, sustained human-to-human transmission has not yet been recorded. H5N1 illustrates this well: human infection does not necessarily imply person-to-person transmission. Public communication in such cases must distinguish between the reality of the biological hazard and the likelihood of a specific outcome. The former might lead to unnecessary public panic, whereas the latter might result in delayed preparedness.
Scientific knowledge will necessarily evolve throughout an outbreak, yet as the use of open-source data and AI-based monitoring systems grows, inaccurate or manipulated data could increasingly influence public perceptions and the identification of threats.
Genomic surveillance has also changed how biosecurity governance is conducted. During the COVID-19 pandemic, sequencing technology enabled the rapid detection of new variants like Alpha, Delta, and Omicron, allowing governments to alter their strategies in response. The distinction between misinformation, disinformation, and genuine scientific disagreement is one of the key challenges of biosecurity governance. Scientific knowledge will necessarily evolve throughout an outbreak, yet as the use of open-source data and AI-based monitoring systems grows, inaccurate or manipulated data could increasingly influence public perceptions and the identification of threats. Addressing this challenge will require standards for evaluating evidence, communicating uncertainties openly, and validating the data that informs decisions about public health measures and underpins AI-based early warning systems.
Building Trust Across Institutions
Trust in biosecurity emergencies should not be confined to one entity. Good governance during pandemics requires cooperation between international organisations, national health bodies, scientific organisations, researchers, media, and digital platforms.
WHO is an international organisation that still bears primary responsibility for coordinating responses to global health emergencies. It provides a platform for collective action through disease surveillance, technical assistance, and international coordination. Yet it can be effective only through collaboration with member states and with transparent reporting. National public health organisations implement international recommendations through domestic policies, surveillance, vaccination programmes, and risk assessment. They derive credibility from scientific independence, transparency, and a separation between technical information and political decisions.
Independent sources of scientific assessment include scientific journals, universities, and researchers. Independent researchers also perform an important function by questioning established views and finding flaws in existing evidence. The media contribute through investigation and accountability.
Policy Foundations for Biosecurity Preparedness
Establishing trust during biosecurity emergencies is a task that can only be achieved through prior planning and preparation. Public health systems should not only ensure sufficient resources for surveillance, laboratory capacity, vaccination, and medical care, but also incorporate communication strategies into preparedness planning.
Public health systems should not only ensure sufficient resources for surveillance, laboratory capacity, vaccination, and medical care, but also incorporate communication strategies into preparedness planning.
Transparency should become integral to public-health decision-making. Governments should clearly explain the basis for their policy decisions, including any uncertainties and the underlying research. Publishing scientific reports and recommendations enables their critical review. Scientific literacy is also crucial for effective communication. Greater public awareness of uncertainty, risk assessment, and evidence evaluation will improve societies’ ability to understand evolving guidelines. Finally, international cooperation is essential, given the borderless nature of infections. The WHO Pandemic Agreement reflects a growing recognition of this need for future outbreaks.
Increased use of such AI-based early warning systems also requires appropriate standards for the governance of open-source data collection and analysis.
AI scrapers can treat all data as a point of reference for studying occurrences, but publicly available information cannot be considered equally credible. It should be verified across multiple sources, with information provenance established, confidence levels assessed, and findings compared against credible public health data. It is important to distinguish between signals originating from social media, blog entries, forums and other sources and information from credible scientific and governmental organisations or peer-reviewed publications, and assign explicit weights to each type of source. Manual oversight of AI-generated results should remain part of the process, particularly when they may inform public health policies or international reporting.
Standards related to regulation and technology need to be put in place for the use of AI in biosecurity surveillance. This should include measures to promote transparency about data sources, validate open-source data, audit for bias and misinformation, and ensure accountability for decisions based on AI outputs. Through such standards, governments can mitigate the risk of misinformation and bias influencing early warning systems.
Conclusion
This divide between misinformation and scientific dissent is perhaps the foremost challenge facing modern biosecurity governance. The COVID-19 experience showed that information management is just as critical as managing the pathogen’s spread. Successful biosecurity governance need not eliminate all disagreement but should ensure that disagreement remains evidence-based and is incorporated into decision-making, scientific communication, data collation, and early warning systems.
Shravishtha Ajaykumar is an Associate Fellow with the Centre for Security, Strategy, and Technology at the Observer Research Foundation.
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.