AI mosquito trap aims to speed up dengue surveillance in Florida

In many parts of the country, November means fewer mosquitoes. Florida is different. The state is facing its largest dengue outbreak in decades, with hundreds of people infected locally rather than during travel.

Researchers at the University of South Florida have developed an AI-powered mosquito trap that could help health officials identify potentially dangerous species faster. In laboratory tests, the system identified mosquito species with more than 98% accuracy. With dengue still spreading in Florida, quicker surveillance could prove useful.

According to the Florida Department of Health’s latest available weekly report, Hillsborough County had recorded 407 locally acquired dengue cases in 2026, the county’s largest outbreak on record. Statewide, officials had reported 524 locally acquired cases through October 31.

The Tampa area has been particularly affected. In October, state health officials confirmed that Hillsborough County was still responding to an ongoing outbreak. Dengue can cause high fever, severe headaches and intense muscle pain. Serious cases can lead to life-threatening complications.

On November 13, Hillsborough County extended its state of local emergency through November 19 as mosquito control crews increased aerial and ground treatments. County officials say surveillance continues to detect dengue virus activity in neighborhoods across the area.

Mosquitoes may be less active in colder parts of the country, but they can persist in warmer states. Florida’s outbreak is a reminder that the risk does not necessarily end with summer. Researchers hope new monitoring technology could help communities respond before mosquito populations become harder to manage.

The system was developed by Digitomy, a startup founded by USF researchers Associate Professor Ryan Carney and Professor Sriram Chellappan. Their work on the technology dates to 2018. The aim is to help mosquito control teams identify insects without spending days examining them under a microscope.

The trap can be hung from a low tree branch or placed on the ground. A light or bait attracts mosquitoes, and a small fan draws them inside. After the insects land on a sticky pad, two built-in cameras photograph them. AI compares the images with a large database to identify the species.

The system can also determine whether a mosquito is male or female and whether it has recently fed on blood. That distinction matters because female mosquitoes bite people and can spread disease. In laboratory tests, the researchers say, the system identified species with more than 98% accuracy.

Faster identification could help guide mosquito control

Traditional mosquito surveillance can take two to three days, since specialists often identify captured insects under a microscope. That work is important, but delays can make it harder for officials to respond quickly during an outbreak.

Digitomy hopes its AI and cloud-based system can identify captured species on the same day. Faster results could help mosquito control teams track where potentially dangerous species are appearing and decide where to focus their efforts.

One species of particular concern in Florida is Aedes aegypti, which is primarily responsible for spreading dengue there. Thousands of mosquito species exist worldwide, but only some transmit particular diseases. Knowing which species are present can help officials make more informed decisions.

There is an important limitation: the trap identifies species that can carry disease, but it does not establish whether a particular mosquito is infected with dengue. Laboratory testing would still be needed to confirm the virus.

The system can send its findings to a mapping dashboard, where mosquito control teams can review activity by location. If traps begin detecting more Aedes aegypti in a particular neighborhood, officials could investigate and consider targeted control measures. That kind of information might help them act sooner, though it cannot guarantee an outbreak will be prevented.

The researchers have added trap data to a private section of their Global Mosquito Observations Dashboard. The platform also includes more than half a million reports from citizen scientists. Digitomy’s system could work alongside an early warning system Carney developed for dengue and West Nile virus, giving public health departments a broader picture of mosquito activity.

Digitomy has tested the technology at USF and worked with the Anastasia Mosquito Control District in St. Johns County, which includes St. Augustine. The researchers are preparing for field validation with Hillsborough County, one of the areas hardest hit by Florida’s dengue outbreak.

The startup is working with USF’s Technology Transfer office to license the AI system to established mosquito trap manufacturers. That approach could allow manufacturers to add the technology to equipment they already make. No consumer launch date or retail price has been announced.

In summer 2026, Digitomy received a $25,000 grant from the USF Research Foundation’s Bull Ring Accelerator to help move the university-developed technology toward licensing and commercialization. The team is also developing an online learning platform where mosquito control professionals and members of the public could practice identifying species using images and AI-generated feedback.

Most people are unlikely to install one of these traps in their backyard soon. Still, if mosquito control departments adopt the system, it could help them detect concerning populations earlier and decide where to intervene. For now, people in warm-weather areas should continue to take basic precautions, especially when traveling.

Empty standing water from flowerpots and other outdoor containers, since mosquitoes can breed in small amounts of water. Use an EPA-registered insect repellent outdoors, and consider wearing long sleeves and pants to reduce exposed skin.

AI-powered mosquito traps could become a useful tool, but their performance in real-world conditions still needs to be tested. The laboratory results are promising. If the system works as intended outside the lab, it could give communities more time and better information to respond to mosquito-borne diseases.