Editorial illustration for HHS develops AI to generate hypotheses on unverified vaccine injury claims
AI Assists VAERS Analysis for Vaccine Injury Insights
HHS develops AI to generate hypotheses on unverified vaccine injury claims
The Department of Health and Human Services is rolling out a new artificial‑intelligence system designed to sift through the flood of vaccine‑injury reports that have never been medically confirmed. Its purpose, officials say, is to help researchers formulate plausible explanations for patterns that emerge in the data, rather than to declare causality outright. Critics have long warned that the national Vaccine Adverse Event Reporting System (VAERS) was never meant to serve as a definitive proof‑of‑harm database; it simply collects raw, unverified submissions from the public.
By feeding those entries into a machine‑learning model, HHS hopes to spotlight clusters worth deeper study while acknowledging the system’s inherent constraints. The move reflects a broader effort to bring more rigor to a field where anecdote often outruns evidence, and it underscores why experts remain cautious about drawing firm conclusions from VAERS alone.
Because these claims are not verified, VAERS data alone can't be used to determine if a vaccine caused an adverse event. "VAERS, at best, was always a hypothesis‑generating mechanism," says Paul Offit, a pediatrician and director of the Vaccine Education Center at Children's Hospital of Philadelphia.
Because these claims are not verified, VAERS data alone can't be used to determine if a vaccine caused an adverse event. "VAERS, at best, was always a hypothesis-generating mechanism," says Paul Offit, a pediatrician and director of the Vaccine Education Center at Children's Hospital of Philadelphia who was previously a member of the CDC's Advisory Council on Immunization Practices. Anybody can report, and there's no control group." Offit says the system only shows adverse events that happened at some point following immunization; it doesn't prove that a vaccine caused those reactions.
CDC's own website says that a report to VAERS does not mean that a vaccine caused an adverse event. Despite this, anti-vaccine activists have misused VAERS data over the years to argue that vaccines are not safe. Leslie Lenert, previously the founding director of the CDC's National Center for Public Health Informatics, says government scientists have been using traditional natural language processing AI models to look for patterns in VAERS data for several years, so it's not surprising that HHS would move toward the adoption of more advanced large language models.
One major limitation of VAERS is that it doesn't include data on how many people received a vaccine, which can make events logged in the database seem more common than they actually are.
Will the new AI tool actually clarify vaccine safety? HHS says it will scan VAERS reports for patterns and then spin out possible explanations. The system is still in development and has not yet been deployed, according to the agency’s inventory.
Because the claims in VAERS are unverified, the database alone cannot prove causation. Paul Offit reminds us that VAERS has always been a hypothesis‑generating mechanism, not a definitive source. Consequently, the AI’s output will be another set of hypotheses, not confirmed findings.
Whether the tool will uncover meaningful links or simply produce more speculation remains unclear. The agency’s 2025 AI inventory lists the project among several use cases, but no performance data are available. In short, the initiative adds a generative layer to an existing reporting system, yet its practical value for public‑health decisions is still uncertain.
Further testing will be needed before the hypotheses can be evaluated against clinical evidence.
Further Reading
- Product Hunt - AI Tools - Product Hunt
- There's An AI For That - TAAFT
Related Reading
AI Tools & Apps Claude Code 2.1.0 launches with smoother workflows, smarter agents for power users
AI Tools & Apps Build a Smart AI Voice Assistant Quickly with Vapi: Step-by-Step
AI Tools & Apps Demystifying AI Workflows: 7 Tools That Boost Transparency and Efficiency
From Archives HHS employs Palantir AI to audit grants for DEI and gender‑ideology compliance
From Archives Firefox adds toggle to disable AI features, matching Edge and Chrome
Common Questions Answered
What is the primary purpose of the new HHS AI system for VAERS?
The AI system is designed to help researchers generate hypotheses about unverified vaccine injury reports by identifying potential patterns in the data. It aims to formulate plausible explanations for emerging trends in adverse event reporting, without declaring definitive causality.
How does VAERS function as a vaccine safety monitoring tool?
VAERS is a national early warning system that accepts reports from anyone, including patients, healthcare providers, and manufacturers, about potential adverse events following vaccination. The system is not designed to prove that a vaccine caused an adverse event, but rather to identify unusual or unexpected reporting patterns that might indicate potential safety concerns.
Why can't VAERS data be used to definitively prove vaccine injuries?
According to experts like Paul Offit, VAERS is fundamentally a hypothesis-generating mechanism with significant limitations. Anyone can submit a report, there is no control group, and the claims are not medically verified, which means the data cannot be used to establish a direct causal relationship between vaccines and reported adverse events.