DARPA is using AI to reduce the modelling time for viral pandemic outbreaks from weeks to days

Speed is prioritized over critical review: AI-generated models are intended to justify interventions even before they can be meaningfully questioned.

The US military is funding artificial intelligence (AI) systems designed to drastically accelerate the modelling of viral outbreaks – a process that normally takes weeks, to be compressed into a few days and then used to guide real-world interventions.

In other words, the faster the model, the less time there is to even question whether the reaction is justified.

This acceleration follows pre-covid DARPA pandemic infrastructure programs that are already documented and aim to convert digital genetic sequences into synthesized viruses and mass-produced mRNA countermeasures within a fixed timeframe.

DARPA’s stated problem: Pandemic models were fragile, opaque, and slow.

According to a Science publication from December 2025:

“As SARS-CoV-2 spread across the globe in 2020, epidemiologists endeavoured to predict its spread – and its deadly consequences. They frequently relied on models that could not only simulate virus transmission and hospitalization rates, but also predict the effects of interventions: masks, vaccines, or travel restrictions.

But aside from being computationally intensive, models in epidemiology and other disciplines can be black boxes: millions of lines of old codebase that are subjected to delicate fine-tuning by operators at research institutions around the world. They don’t always provide clear guidance. ‘The models that are used are often quite fragile and inexplicable,’ says Erica Briscoe, who was the program director for the Automating Scientific Knowledge Extraction and Modelling (ASKEM) project at the Defense Advanced Research Projects Agency (DARPA).”

The Defense Advanced Research Projects Agency (DARPA)’s own program director thus acknowledges that the models that guided responses during the covid era were vulnerable and difficult to interpret.

This means that after covid, there will be no attempt to slow down or restrict model-driven policy.

Rather, the aim is to make the same decision-making machine run faster.

There is “real potential” to accelerate modelling during an outbreak, says Mohsen Malekinejad, an epidemiologist at the University of California San Francisco who helped evaluate the ASKEM products. “In a pandemic, time is always our biggest constraint. We need to have the information. And we need to have it fast,” he says. “We simply don’t have enough data-literate modelers for every single occurrence or every different type of public health need.”

The program: AI-generated outbreak models on demand

“The program, launched in 2022 with a budget of US$29.4 million, aims to develop AI-based tools that make modelling easier, faster and more transparent.”

DARPA has funded an infrastructure that standardizes and accelerates the modelling of outbreaks. The focus is on speed, reproducibility and usability by non-specialists, so that politically relevant models can be generated quickly – even if the underlying assumptions are incomplete or controversial.

How it works: Technical articles & notebooks → Equations → Models

“The program’s AI tools automate this programming and allow researchers to construct, update, and combine models at a higher level of abstraction.”

By removing much of the technical friction involved in model building, these tools facilitate the creation of outbreak models that carry institutional weight – even when the scientific basis is thin or uncertain.

“ASKEM teams developed AI systems that can absorb scientific literature……. and extract the equations and knowledge needed to create or update a particular model.”

Scientific literature is directly converted into reusable model components, enabling machine-interpreted research results to be quickly incorporated into decision frameworks.

“An ASKEM project developed a method to process these notebooks, extract the underlying mathematical descriptions, and translate them into a model.”

Informal considerations and exploratory notebook work can thus be transformed into usable models at high speed, thereby shrinking the gap between preliminary thinking and authoritative results.

Intervention-focused modelling

“The resulting model integrated the different transmission and seasonal patterns of the viruses and also assessed the effects of interventions such as mask-wearing and testing.”

The system is designed to evaluate intervention scenarios in parallel with disease dynamics and to embed policy considerations directly into the modelling process.

“Testers were asked to model the impact of a vaccination campaign on the cost of hospitalizations for hepatitis A among homeless drug users in a federal state.”

These tools are geared towards applied government issues – costs, target groups and campaign impact – and not purely descriptive epidemiology.

The speed claim: 83% faster

“In the final results, the testers found that the ASKEM tools were able to create models 83% faster compared to standard modelling workflows.”

The modelling process is fast enough to fit into political and media timeframes, thereby reducing the possibility of external review before results are translated into action.

“They were able to create practically usable models for several problems within a 40-hour work week.”

Once speed is no longer the limiting factor, the pressure shifts towards rapid implementation rather than careful validation.

“Transparency” as an internal signal of trust

“Due to the increased transparency of the ASKEM models, the testers also found that decision-makers had more confidence in the results of ASKEM than in those of traditional models.”

Here, “transparency” functions less as a protective mechanism and more as a way to increase trust among decision-makers.

By making the models understandable enough to pass internal audits, the system reduces friction within institutions and enables those in charge to act faster, while still including unresolved uncertainties in the results.

Intended users: Health, defence, and intelligence agencies

“DARPA is working to find agencies or programs within the health, defence, and intelligence communities that would like to use ASKEM.”

Outbreak modelling is positioned as a permanent national security capability and integrated alongside defence and intelligence functions, rather than being treated as an ad-hoc public health measure.

Conclusion

DARPA is building a system that transforms scientific literature, assumptions, and exploratory analyses into outbreak models – fast enough to guide interventions in near real time.

If speed is considered the primary constraint, the window of opportunity for review, objection, and meaningful questioning inevitably collapses before these models can be used to justify action.

 

yogaesoteric
January 31, 2026