In September 2026, OneSTOP was featured in seven posters at the Neobiota 2026 conference. Each of them presents a different aspect of the project and how it helps minimise the spread and impact of invasive alien species (IAS) in an innovative way. They will be available online until 31 March 2027 on Neobiota’s online poster platform and can also be accessed here as a series of factsheets.  

Detection of invasive plants and insects with novel technology

Some of the technologies used by OneSTOP are AI-powered camera systems that can detect invasive species at continental scale: CamAlien and AMI traps. CamAlien is a camera mounted on a vehicle that photographs roadsides, and its images are analysed by a deep-learning pipeline built with Pl@ntNet. So far, more than 13,000 invasive plant records gathered by this system have already been published to GBIF. The OneSTOP team is also working on detecting roadkill from the images. The other tool used is an AMI trap, which is an automated light trap for night-flying insects. This tool has been placed in botanical gardens in 14 countries, where it has detected the damaging box tree moth and several other invasive insects. Learn more about this poster here. 

Engaging stakeholders in the detection and monitoring of alien species through the OneSTOP Living Labs

OneSTOP has established five Living Labs across Europe, bringing together local stakeholders to co-create innovative IAS detection and monitoring technologies. So far, they involve more than 100 people from about 50 organisations across business, government, research, civil society and the environment sector. The Living Lab core groups meet two or three times a year. They have already deployed detection technologies such as AMI traps, airDNA samplers and CamAlien, and conducted BioBlitz events. In the future, they plan to engage citizens more actively, test the OneSTOP Alert System, and pursue other activities. Learn more about this poster here. 

Sentinel gardens for the early detection of invasive species using Shiny apps

The next poster stays on the topic of engagement, presenting OneSTOP’s sentinel gardens and web apps. While gardens are the place where invasive plants often start spreading, researchers find them hard to access. The idea behind the sentinel gardens is to engage gardeners, who often notice which plants are spreading, so they can act as an early-warning network. In parallel, the PlantAlert Shiny app allows citizens to identify and report invasive garden plants, with all records openly available on iNaturalist. The app builds on earlier national projects across Europe and existing platforms such as iNaturalist and Pl@ntNet. Another Shiny app for invasive flatworms has also been launched across the OneSTOP Living Labs. Learn more about this poster here. 

Using large language models for creating invasive species management plans 

This poster presents the results of a study that tested whether free AI chatbots such as ChatGPT, Copilot and Gemini can write useful management plans for invasive alien species. The models produced more than 100 plans for three invasive alien species, chosen by the UK Living Lab: rhododendron, demon shrimp and American mink. Their overall structure was plausible, but they varied in length, and some used unexpectedly militaristic language. More importantly, practical details such as cutting heights and herbicide timing changed between runs and sometimes contradicted each other. The conclusion reached is that while AI can help draft local plans, they must be checked carefully against best practice before anyone acts on them. Learn more about this poster here.

Accelerating EICAT assessments with AI and evaluating human-AI agreement 

The Environmental Impact Classification for Alien Taxa (EICAT) is the IUCN global standard for classifying the severity of environmental impacts caused by animals, fungi and plants living outside their natural range. It is a key tool for prioritising alien species that could lead to the most harmful environmental impacts, helping to make the best use of resources to prevent or limit their negative consequences. However, preparing an EICAT assessment normally requires analysing numerous papers by hand. To accelerate this process, OneSTOP researchers built an AI agent that extracts impact evidence from papers and classifies it into valid EICAT categories. The papers assessed by the AI agent were also checked by human experts; however, human annotators disagreed with each other substantially on this task. This motivates the need for a human-in-the-loop to make a final decision based on what is extracted by the AI model. Learn more about this poster here.

Prioritising impact of IAS on European threatened species

The EU Biodiversity Strategy aims to cut in half the number of Red List species threatened by invasive species by 2030. OneSTOP helps achieve this goal by proposing a semi-automated, reproducible workflow that converts the IUCN Red List threat information into quantitative scores. Those numbers are combined with distribution models of both native and invasive species to produce a cumulative impact score, called the CIMPAL-RL index. The workflow has already been tested in the Iberian Peninsula, using 14 pairs of native birds or mammals and the invasive species affecting them. It works by creating prioritisation maps, which will be fed into conservation planning tools and show where management should focus, as well as which areas could serve as refuges for threatened species. Learn more about this poster here.

Accelerating invasive alien species detection with online data mining

The last poster is based on a study that tested whether online activity can act as an automated early-warning system for 88 invasive species of Union concern between 2016 and 2025. It covered Wikipedia page views and posts on Flickr, YouTube, Facebook and iNaturalist. The result was that online activity reflected real invasions weakly and unevenly compared with official occurrence records. For example, Wikipedia performed much better than Flickr, and visible, charismatic animals got far more attention than plants. In addition, spikes in activity matched new invasions only about 40% of the time. Overall, the conclusion reached was that this approach had limited potential for systematic, large-scale early detection. Learn more about this poster here and read the policy brief on the topic here.