The COVID-19 pandemic was many things at once: a health emergency, an economic shock, and a profound test of how societies share information. Perhaps one of its most unexpected lessons was that a virus spreads not only through the air and human contact, but also through the flow of words, images, and claims across digital spaces. Public health advice, scientific research, political rhetoric, and social media became tangled together in ways that made the outbreak harder to manage. Decisions had to be made quickly, often with incomplete evidence, while the same evidence was being debated publicly, sometimes in real time. At the same time, misinformation about the virus, vaccines, and treatments was moving faster than any official announcement. The World Health Organization called this an “infodemic,” a situation where true and false information are mixed so thoroughly that people struggle to know what to believe. This is not just a communication problem; it is a biosecurity problem. If people cannot distinguish reliable guidance from rumor, they may ignore precautions, delay treatment, or lose trust in public institutions. The challenge becomes even more complicated with the rise of artificial intelligence. AI is now being used to scan vast amounts of open-source information—social media posts, news reports, blogs, forums—to detect early signs of outbreaks and support early warning systems. That means the same digital environment that spreads misinformation is also being used as a tool for monitoring threats. So a crucial question emerges: how do we tell the difference between honest scientific communication, genuine scientific disagreement, misinformation, and deliberate disinformation, especially when machines are helping us interpret the noise?
During the early months of COVID-19, scientific advice shifted as understanding evolved. At first, officials emphasized surface transmission and respiratory droplets; later, airborne transmission came to be seen as far more important. Recommendations on masks, ventilation, and social distancing changed accordingly. For many people, these shifts looked like confusion or even dishonesty, not the normal process of science correcting itself. Scientific knowledge is supposed to evolve through hypothesis, testing, peer review, and replication. In an outbreak, different researchers may legitimately disagree because the evidence is incomplete. The debate about airborne transmission is a good example: aerosol scientists argued early on that the virus could linger in the air, while official guidance remained focused on droplets and surfaces. Both sides were working with limited data, and neither was trying to deceive the public. Over time, the evidence became clearer, and guidelines changed. That is how science works. But in a crisis, this kind of honest uncertainty can be easily mistaken for incompetence or conspiracy. The problem is made worse by the fact that political leaders, media outlets, and ordinary users all have incentives to amplify certain claims. Some people deliberately spread false information, but many others simply share what they believe is true, adding another layer of confusion. The line between misinformation and dissent is not always obvious. A scientist questioning official guidelines is not necessarily spreading misinformation; they may be doing exactly what scientists are supposed to do. Yet in the digital ecosystem, that dissent can be pulled out of context, exaggerated, or weaponized by bad actors. Distinguishing between productive disagreement and dangerous misinformation is one of the core governance challenges for biosecurity today.
The rise of AI in biosecurity surveillance makes this challenge more urgent. Modern AI systems can process enormous amounts of publicly available data, looking for patterns that might signal an emerging outbreak before traditional systems catch it. This is genuinely promising. It can help detect unusual disease clusters, monitor rumors, and give health authorities a head start. But there is a serious risk: AI tools do not automatically know whether the information they are analyzing is true. They treat data as data)Skip. If a social media post claims that a strange illness has appeared in a certain town, an algorithm might flag it as a signal. That could be a real early warning, or it could be a hoax, a misunderstanding, or a deliberately planted piece of disinformation. If the system cannot tell the difference, it may produce false alarms, distract resources, or even miss real outbreaks because the noise is too loud. The quality of the open-source information feeding these systems matters just as much as the sophistication of the algorithms. During the pandemic, false claims about the origin of the virus, miracle cures, and harmful vaccine side effects all spread rapidly online. If AI-powered early warning systems absorb such content without careful verification, they may end up amplifying the very misinformation that makes emergencies worse. This is why information governance must be part of AI development. It is not enough to build faster or smarter detection tools; we also need standards for evaluating sources, assigning confidence levels, and distinguishing between credible reports and rumor. Human judgment must remain in the loop, especially when AI outputs are used to inform public health decisions or international reporting. Otherwise, we risk building a surveillance system that is highly efficient at spreading confusion.
The communication challenges of biosecurity are not limited to COVID-19. The mpox outbreak of 2022 showed how deeply context and trust are intertwined with public health messaging. Because the disease initially spread within certain sexual networks, officials had to warn people without fueling stigma or discouraging testing and treatment. Even scientifically accurate information can do harm if it is delivered without sensitivity to the social realities of the affected communities. If people feel blamed or judged, they are less likely to seek care or cooperate with contact tracing. This is a reminder that communication is not just about transmitting facts; it is about building relationships and respecting the people receiving the information. The same lesson applies to highly pathogenic avian influenza, or H5N1. As of now, the virus has caused some human infections, but sustained human-to-human transmission has not been recorded. Public communication must strike a delicate balance. If officials emphasize the severity of the biological hazard without explaining the low likelihood of a pandemic, they may cause unnecessary panic. But if they downplay the risk too much, they may leave communities unprepared. The challenge is to be honest about what is known and unknown, and to communicate probabilities rather than certainties. Genomic surveillance has also changed the picture. During COVID-19, rapid sequencing allowed scientists to identify new variants and respond accordingly. This kind of science is immensely valuable, but it also depends on public trust. People need to understand why monitoring viruses at the genetic level mattershare and how their data and samples are being used. As open-source data and AI-based monitoring become more common, the risk of inaccurate or manipulated information influencing threat identification grows. Scientific knowledge will always evolve during an outbreak, but we need stronger standards for how evidence is gathered, how uncertainty is communicated, and how the data that informs AI systems is validated.
Trust during a biosecurity emergency cannot rest on any single institution. It has to be distributed across a network of organizations, each playing a different but complementary role. The World Health Organization coordinates global response efforts, provides technical assistance, and monitors disease trends. But it can only be effective if member states report transparently and if its recommendations are seen as scientifically independent rather than politically motivated. National public health agencies translate those recommendations into local action, through vaccination campaigns, surveillance, and risk communication. Their credibility depends on openness and on a clear separation between technical advice and political pressure. Independent researchers, universities, and scientific journals also play a vital role. They can challenge received wisdom, catch mistakes, and push the evidence in new directions. That kind of institutionalized doubt is essential; but it only works if it happens in good faith and with rigorous standards. The media, too, are part of the trust ecosystem, both as watchdogs and as translators of complex information. Building this kind of trust requires preparation before a crisis hits. Communication strategies should be embedded in public health preparedness planning, not added as an afterthought. Transparency should be a default: governments should explain the evidence behind their decisions, acknowledge uncertainties, and publish the scientific reports that guide policy. Scientific literacy also matters. If the public understands how evidence evolves and why recommendations change, they are less likely to see every shift as a betrayal. International cooperation is just as important. Infectious diseases do not respect borders, and neither should our systems for sharing information and coordinating response. The WHO Pandemic Agreement reflects a growing recognition of this need, but it must be backed by real commitment from all countries.
In the end, the greatest challenge for biosecurity governance may be learning to live with disagreement. We do not need to eliminate scientific dissent; in fact, we need it. But we do need to ensure that disagreement remains evidence-based, transparent, and protected from manipulation. That means creating an information environment where uncertainty can be expressed without being distorted, where AI systems are designed to distinguish credible sources from noise, and where public health decisions are made visible and explainable. It also means recognizing that misinformation is not simply a problem of ignorance. It is a symptom of deeper social and political dynamics, including distrust of authority, fear of change, and the human tendency to seek information that confirms what we already believe. Addressing it requires more than fact-checking; it requires building relationships, listening to concerns, and making people feel included in the decisions that affect their lives. The lesson of COVID-19 is that managing information is just as important as managing the virus itself. Future emergencies will be shaped not only by biology and technology, but by how well we communicate about them. If we can learn to handle uncertainty honestly, to challenge misinformation without silencing dissent, and to build early warning systems grounded in trust and verification, we will be better prepared for whatever comes next. But this is not a task for scientists or governments alone. It is a shared responsibility for all of us.