DEEP DIVE – It’s the opening act in a potential public health nightmare: a chicken dies on a farm, for no apparent reason; another perishes at a farm hundreds of miles away; it takes time for the farm owners to notice, more time for tests to be conducted and different anomalies connected, and before the diagnostics are complete, the damage is done – the first wave of a bird flu pandemic has broken.
Beyond natural outbreaks, there are also concerns involving deliberate acts: This week the Department of Justice charged three Chinese nationals with smuggling biological materials into the U.S.; and in June two Chinese researchers were charged with trying to smuggle a fungus into the U.S. that can devastate grain crops.
Some experts are imagining a world in which technology is harnessed to ensure that such biosecurity nightmares don’t happen – or are dealt with much faster and more effectively.
“What we're promoting is a system that can look at things more holistically and on a much larger scale,” Robert Norton, a professor of veterinary infectious diseases and coordinator of national security and defense projects at Auburn University, told The Cipher Brief. “The system is designed to fill gaps in biosurveillance, looking for disease outbreaks, whether they be naturally occurring or induced through bioterrorism.”
That proposed system has a name – BISR, for Biosurveillance Intelligence, Surveillance, and Reconnaissance – and its backers believe it would revolutionize the field of biosurveillance. The core concept is that sophisticated sensors and other tools used by the U.S. Intelligence Community (IC) can be leveraged to improve detection, and that artificial intelligence can be deployed to help fast-track diagnosis. The chicken-farm example is only one scenario; responses to a COVID-19-like outbreak or acts of bioterrorism would be improved as well.
Norton, Daniel Gerstein, a senior policy researcher at RAND, and Cris Young, professor at the College of Veterinary Medicine at Auburn, co-authored an article last year arguing that the creation of a BISR system was “a national security imperative at the crossroads of technology, public health, and intelligence.” The BISR, they wrote, “would be designed to address two mission-critical requirements for biosurveillance: rapid detection and predictive analysis.”
They have taken their plans to Capitol Hill – specifically, to the House Permanent Select Committee on Intelligence, where they say they have received “good reviews.” The Select Committee wouldn’t comment on the BISR proposal itself but in a statement to The Cipher Brief, a spokesperson said that “The Committee continues to explore various biosecurity initiatives and programs to ensure that the U.S. is postured sufficiently to combat and prevent any future biosecurity threats that could cause widespread harm.” The statement went on to say that the Committee is working with the Office of the Director of National Intelligence (ODNI) “to establish an Office of Intelligence within the U.S. Department of Agriculture to address threats to U.S. agriculture.”
The threats are clear, to agriculture and beyond. The U.S. remains vulnerable to biologically driven disruption – be it from another COVID-like pandemic, an outbreak of bird flu that reaches humans, or bioterrorism. Anxiety over the latter has grown as experts worry that AI may be used to create dangerous biological pathogens.
At last year’s Cipher Brief Threat Conference, Jennifer Ewbank, a former CIA Deputy Director for Digital Innovation, warned of “the application of AI in biological weapons by unsavory actors.” And a 2024 report from the Johns Hopkins Center for Health Security said that the same AI capabilities that might produce medical breakthroughs could – inadvertently or otherwise – lead to the creation of deadly pathogens. AI models may “accelerate or simplify the reintroduction of dangerous extinct viruses or dangerous viruses that only exist now within research labs,” the report found.
How prepared is the U.S. to counter such threats? And might a technology-driven “BISR” system revolutionize biosurveillance, as its backers contend?
Save your virtual seat now for The Cyber Initiatives Group Winter Summit on December 10 from 12p – 3p ET for more conversations on cyber, AI and the future of national security.
How a “BISR” might work
The crux of the case for a BISR system is twofold: first, that an array of sophisticated data-gathering tools – drones, satellites, hyperspectral sensors and others – can be mobilized to track biosecurity anomalies; and that trained AI models would analyze the data that the system collected. The system’s architects envision a BISR “dashboard” that provides first responders and decision makers in government, the military and business near-real time insight and analysis.
It’s a high-tech effort to gather clues – a change in a community’s waste water, a spike in the sales of certain medications, even the breathing or social behavior of animals – and assess their meaning more rapidly than current systems allow.
“Our system is agnostic,” Norton said. “It doesn’t matter whether it’s a natural disease outbreak or a terrorism event, it’s looking for those changes and then being able to rapidly detect them and rapidly alert the individuals that are responsible.”
To expand on the chicken-farm scenario: at the moment, one animal’s death might lead a farm worker to call the company veterinarian, the veterinarian would take samples, the farm would look at the flock as a whole, and samples would be brought to laboratories for tests. Ultimately the case might go to a national lab to determine whether avian influenza or another condition was present.
Public health officials say the current system works – but can be slow. Advocates for the BISR system say it would at minimum improve the speed of response, gaining valuable time to determine not only whether a virus was present, but also how it might be circulating in the broader environment. Sensors in and around the poultry houses would track not only a dead chicken, but also the emissions and even behavioral anomalies within the flock – “pattern-of-life” behavior, as the experts say. Any anomaly would be flagged and the system “tipped off,” as Auburn’s Cris Young put it, to alert sensors on other farms.
“The sensors would tip and cue other sensors that would then take a larger look at the larger area or even a state,” Young told The Cipher Brief, “to determine if those signatures coming off of that one particular house that's affected are similar to things happening in other houses.”
Given the sheer volume of data generated by a BISR system, AI models would be used to rapidly assess the data – and check anomalies against specific pathogens.
BISR’s proponents say a similar approach could be taken with viruses among humans, providing more rapid early-warning mechanisms and analysis.
“Advances in sensor capabilities, coupled with the use of AI platforms, provide new capabilities that could be applied to the detection of biological events in the early stages of an outbreak,” the authors of the BISR article wrote. “The concept would provide new tools for early detection, response, mitigations, and ultimately, recovery from an outbreak.”
Are you Subscribed to The Cipher Brief’s Digital Channel on YouTube? There is no better place to get clear perspectives from deeply experienced national security experts.
The tools of a BISR system
The system’s architects say most of its high-tech elements already exist – sensors in place on poultry farms or in public spaces, and various tools of ISR (intelligence, surveillance and reconnaissance) that are currently used across the IC. These might include multispectral and hyperspectral sensors, and many airborne assets – balloons, drones, aircraft and satellites – that have been used to detect concentrations of gases for national security purposes. The International Space Station, for instance, regularly uses hyperspectral imaging to map the earth’s surface, and the Department of Defense uses hyperspectral imaging for several purposes – including detection of chemical and biological hazards.
Norton cited the example of the IC’s use of satellite imagery to monitor concentrations of nitrate in Afghanistan – because high levels of nitrate often indicated the presence of bomb-making facilities. Nitrate is also a component found in animal waste – and so in the public health example, he said, satellite imagery could be used to monitor levels of nitrate and other compounds on a farm.
Ultimately, BISR’s proponents believe the system could also be used to monitor the volatilome (essentially, what humans and animals breathe out) of people at airports or stadiums or other crowded environments, and alert public health officials about anomalies in the data. Young described a scenario in which international arrivals at Atlanta’s Hartsfield Airport – the nation’s busiest – would be watched by hyperspectral sensors to detect anomalies in respiration.
“We might have sensors set up in multiple places as [people] disembark from their flight,” Young said. “There might be several places to take a different scan with multiple sensors, and we might be able to say with some certainty, this person is infected with let's say COVID, and this person is actually shedding the virus.”
The hope is that any anomaly – be it on a chicken farm or at a crowded airport – would tip the system to sweep up other relevant information: Have ER visits spiked in a community? Does social media from that community suggest related anomalies? And so forth. Ideally, a dangerous pathogen would be flagged and identified before it leads to a pandemic, or an act of bioterror would be detected at the earliest possible moment.
Michael Gates, CEO of GDX Development, a company that bills itself as “solving very complex national security challenges,” says he joined the BISR effort “from the technology side of the equation.” GDX has worked previously with the U.S. Special Operations Command. Gates says the key to BISR’s success will involve “sensor fusion” – the linking of a range of data-gathering mechanisms.
“If you think about the world of the Internet of Things, everything's a sensor, and there's not very many systems out there that have the ability to collect off of all of those sensors, bring that data payload in, and then push it into a single pane of glass that can be used for military operations, for intelligence sharing or more tactical things,” Gates told The Cipher Brief.
In the chicken farm example, Gates envisions “sensor fusion” ranging from a hyperspectral scan to “available drone assets” and ultimately “zeroing in down to sensors such as temperature, air purification, even cameras monitoring chicken behaviors.”
Once a problem has been identified, Gates said, “you can use open-source intelligence and other things to mine, let's say, a Reddit form for these things – is anybody talking on the internet about their chicken coops having issues? – and so on, for whatever the issue is.”
“There's already enough sensors out there,” he added. “The data is there. What's happening is that information's not being shared. It's not being centralized, meaning we're getting delayed responses...Nobody has a holistic picture right now on biosurveillance.”
In the early stages of a crisis, the BISR might do a lot of work before humans are engaged, though the Auburn professors stress that the system aims only to provide experts a head start, rather than cut them out of the proverbial “loop.”
“We support human-in-the-loop artificial intelligence systems,” Young said. “We want there to be a person that has to look at this screen at some point and say, okay, I understand what's going on here. Maybe that happens within minutes of an anomaly occurring, but regardless, at some point a person needs to decide, Yes, that's what this is, or No, we need further information.”
Sign up for the Cyber Initiatives Group Sunday newsletter, delivering expert-level insights on the cyber and tech stories of the day – directly to your inbox. Sign up for the CIG newsletter today.
The challenges
Norton and Young say they have presented their plans to the House Select Committee and are prepared to do the same to the Defense Advanced Research Projects Agency (DARPA). They believe their system can be 80 percent complete in three years and fully functional in five. As for costs, they say the first two years would require a budget of $10 million, and that the system’s operating costs would eventually be $300 million annually. They argue that billions of dollars have been spent in the biosurveillance domain, and that the BISR would be a major upgrade over existing capabilities.
It may sound like a no-brainer – the smart use of technology to guard against myriad biosecurity threats – but questions abound about BISR and its future. And many of the hurdles to its implementation involve, in one way or another, the human element.
Just as the Intelligence Community has struggled at times to share information and assess national security risks, the government architecture in biosurveillance is complex and often siloed. A host of agencies share responsibility for the nation’s biosecurity – the Centers for Disease Control (CDC), the Health and Human Services Department (HHS), the National Institutes of Health (NIH), the Department of Homeland Security (DHS) and the Department of Agriculture (USDA), to name a few. Experts say they don’t always communicate effectively with one another – and that states don’t always share critical information effectively with the federal government.
Dr. Tom Inglesby, Director of the Johns Hopkins Center for Health Security, noted that in the most recent bird flu outbreak in the U.S., some states wanted to handle the information and response without involving the federal government.
“They weren't even very interested in USDA at times,” Inglesby told The Cipher Brief. “So they said, we'll handle this on our own and we'll let you know. Meanwhile, CDC has to wait for states to bring them the data and information. They don't have command authority to say you must deliver it. It's a voluntary basis.”
Norton says the BISR developers are hoping to partner with one “Mother Ship” agency within the IC – he wouldn’t say which one – because the IC controls the government’s most sophisticated satellites and other data-gathering systems. He also said that while the system involves high-tech elements and the building of the BISR “dashboard,” technology isn’t the primary hurdle.
“Biosurveillance is not a technology problem, but rather a permissions and authorities problem,” Norton said. That might involve permission to use a Pentagon satellite for biosecurity purposes, he said, or agreement from a major industrial farm to share its data or house sensors on its property.
Inglesby said that transparency and information-sharing would be critical for a BISR-like system to work – and that in the case of the chicken farm example, key stakeholders might be unwilling to cede control of the analytical process to a BISR “dashboard.”
“You have the farm owner who will want to make his or her own assessment, you have local government that may not want outsiders coming in and making a determination for them, and you might have unwillingness even at the federal level to do this,” Inglesby said. “You’re going to need an across-the-board buy-in that we haven’t always seen.”
There are also questions about technical implementation. In the Atlanta airport example, Norton acknowledged that even a highly sophisticated hyperspectral sensor wouldn’t be able to detect, say, COVID-19, unless passengers were directed to a discrete area close to the sensors – and here again, permissions would be needed to install such sensors. The post-COVID atmosphere has suggested less public appetite in the U.S. for intrusive screening, not more. The House Select Committee, in its statement to The Cipher Brief, included a reference to “ensuring any proposal balances privacy and the need to avoid the abuses of the COVID-19 period.”
Inglesby also stressed the importance of transparency on the global stage when it comes to public health crises. In the early days of the COVID-19 pandemic, China failed to share the detailed casework of its first 500 patients in the “ground-zero” city of Wuhan – and more than five years later, it still hasn’t done so.
“In Wuhan, the data was very available, there were a lot of people dying, but the data was covered up,” Inglesby said. “And so even if you had installed the most sophisticated systems, if they're being run by people who don't want to share that information, it's not going to change anything.”
Some early-warning biosurveillance systems are already in place, in the world of what’s known as “Syndromic Surveillance” – and experts say many have worked well.
The CDC’s BioSense platform gathers health-related data from hospitals and clinics to detect potential outbreaks or bioterrorism events. As a part of BioSense, "Sentinel Alerts" are generated when reports involve high-concern viruses or diseases. In the case of influenza (the human variant), alerts are triggered when more than 3 % of ER visits are for the flu. Globally, satellites have been used to track dengue fever outbreaks by measuring water levels in the jungle. And wastewater surveillance systems exist to check on levels of bacteria or viruses.
A less positive precedent is the BioWatch program, which was created by DHS in 2001 and billed as "the nation's first early warning network of sensors to detect biological attack." The system tracks the air supply using Environmental Protection Agency air filters, and sends information to the CDC and – if warranted, to the FBI. The system has been blamed for generating dozens of false positives, and in an audit reported by the Associated Press in 2021, BioWatch was said to have failed in detecting known threats.
Norton told The Cipher Brief that today’s technologies are sophisticated enough to ensure that BISR would operate at a higher level than BioWatch. He added that rigorous standards in the AI models would “prevent AI hallucinations” that could cause false positives – or worse, false negatives.
And Inglesby was quick to note that any improvements in early warning and diagnostics would be welcome.
“There is no single system in the country, and people have been talking about building stronger biosurveillance for a long time,” he said. “Anything you can get done in this space would be super-valuable, assuming the costs aren’t prohibitive and you get the buy-in to use this information wisely.”
Read more expert-driven national security insights, perspective and analysis in The Cipher Brief
