Environment | Promising | SDG14 | SDG15 | United States


Share this post

1. General


SDG 14: Life Below Water

SDG 15: Life on Land



2. Project Details

Company or Institution

Wild Me



General description of the AI solution

Wildbook (wildbook.org) is an autonomous cloud-based, data-driven, Artificial Intelligence-enabled system that turns crowdsourced photographs into information about animals. Wildbook combines heterogeneous wildlife data (especially mark-and-recapture “tagging”, social, genetic, etc.) with machine learning techniques to enable large-scale (regional and global) collaborative wildlife studies in the fields of population biology, animal biometrics, social ecology, molecular ecology, toxicology, and more. Wildbook is especially effective where 1) collaborative and citizen-science data can be incorporated to dramatically increase data volume; 2) computer vision and artificial intelligence (AI) can be harnessed to classify species and identify individual animals by their markings (“photo-identification”) in abundantly available imagery, speeding data curation and increasing both volume and accuracy; and 3) collaboration within and between disciplines can yield new discoveries through data science (e.g., comparative population assessment to discover environmental stressors and other drivers of population declines) and novel machine learning research.

Using Wildbook, the occasional, years-long projects of the past become inexpensive, continuous monitoring efforts supported by public engagement and collaborative research teams, allowing the impact of conservation strategies and the health of wildlife populations to be quickly evaluated and compared at greater scope, speed, and scale than ever before. By providing unified tools and software interfaces for biologists, ecologists, conservation managers, computer scientists, and statisticians, Wildbook is the foundation for advanced studies to better understand and protect endangered populations, directly applying multi-modal, multi-stage machine learning to support new analyses ranging from molecular studies to global population assessments.




Wild Me

3. Aspects

Excellence and Scientific Quality: Please detail the improvements made by the nominee or the nominees’ team or yourself if your applying for the award, and why they have been a success.

Imagery from tourism and scientific studies is one of the most abundant sources of wildlife data. If these images could be widely obtained, rapidly analyzed and combined with related data (e.g. location, date, behavior), then ecologists could benefit from larger and broader data sets to develop data-driven strategies to measure, understand and reverse terrestrial and marine species declines (SDGs 14/15).

Wild Me (wildme.org) is an NGO of software and machine learning engineers. Our primary project “Wildbook” blends AI, structured data, collaboration security, and citizen science. Wildbook provides a web-based, technical foundation (retrainable AI, APIs, user interface) for wildlife researchers to:

-track individual animals in a population using natural markings automatically detected and matched with AI to build mark-recapture statistical models (e.g., tracking abundance)

-scalably engage citizen scientists in data collection

-build a collaborative, distributed network of researchers freely using structured data and common analytical techniques to develop local and regional conservation strategies

Wildbook includes a multi-modal, multi-species retrainable AI server, which provides a multi-stage, deep learning-based pipeline (PyTorch) for finding one or more animals of one or more species in photos and then routing each detected animal to individual ID algorithms (multiple ML techniques employed), allowing biologists to review matches in support of population analyses. Wildbook supports a pluggable approach to integrating other machine learning (e.g., new ID algorithms emerging from academia or competitions) and serves as the easiest way to get new AI techniques into the hands of wildlife researchers in the field.

Wildbook is open source and supports collaborative mark-recapture, molecular ecology, and social ecology studies. Relevant scientific publications from our AI-based work can be found at:


More information about Wildbook’s AI approach can be found at:


Scaling of impact to SDGs: Please detail how many citizens/communities and/or researchers/businesses this has had or can have a positive impact on, including particular groups where applicable and to what extent.

Since its first instance for whale sharks (Rhincodon typus), Wildbook has grown from a single software installation with online, integrated computer vision into a set of multi-species research platforms for globally-distributed marine and terrestrial research. Wildbook has been cross-applied to tiny weedy seadragons in Australia, African wild dogs in Botswana, snow leopards in Asia, and right whales in North America. A list of Wildbook-based projects can be found here:


Each Wildbook is a cost-effective, multi-user platform to support local wildlife researchers with AI to rapidly process imagery and track endangered populations, especially where migration can bring the same individuals across borders and data sets, requiring collaboration and data standards in science and supporting necessary regional approaches to protection of marine and terrestrial species. Each Wildbook has a common setup, codebase, and refined AI development pathway and pipeline. Our nonprofit engineering staff supports our research community in a transparent discussion and feedback forum:


Wildbook is used by almost 900 active wildlife researchers in tracking over 188,000 individual animals for multiple species across the globe as well as for new AI research in academia. Over 15,000 members of the public have been engaged in data collection, and we are still rapidly expanding our species and community support, developing new Wildbook prototypes and new techniques in human-AI interaction workflows. Scientific output is already significant, with studies ranging from local population analyses to global, collaborative analyses of habitat and biology.


With many species bearing individually identifiable marking and scarring, Wildbook has large potential for growth and impact under SDGs 14/15, but further investment is required to expand its scalability, keep support cost-effective, and push the boundaries of AI for wildlife research.

Scaling of AI solution: Please detail what proof of concept or implementations can you show now in terms of its efficacy and how the solution can be scaled to provide a global impact ad how realistic that scaling is.

With only 8 staff, Wild Me transparently supports almost 900 wildlife researchers across the globe in tracking over 118,000 individual animals from 50+ species. Our technology is designed to scalably put AI into the hands of local researchers who best understand the local conservation actors, issues, and environment. Wildbook provides them with the tools (especially AI) and scalability to rapidly develop and iterate locally effective conservation strategies justified by data-driven analyses.

As an open source platform with distinct user interface and AI components, Wildbook is a software ecosystem that supports application to new species, application to new domains (e.g., AI for anti-trafficking of tigers, shark fins), advancement in multi-stage human-computer interaction using AI, and development of fundamentally new techniques in AI for computer vision (especially as it builds novel, collaborative data sets and is applied to new species with new and different markings). Our open source software is actively being developed here:


And our extensible, multi-stage AI platform is technically described here:


And in more specific detail:


Scaling Wildbook is financially and technically challenging. New species can cost-effectively reuse existing Wildbook infrastructure but require some custom training of AI (and therefore relatively experienced AI practitioners) and present subtle, new challenges in individual ID (e.g., rapid accumulation of scarring on gray whales can change their visual presentation). Generally, conservation funding is more structured toward small support for local conservation efforts and not in support of global-scale, dedicated, professional (and expensive) AI teams like Wild Me. As a nonprofit, Wild Me must hire and retain very talented engineers (competing with Facebook, Google, etc.) while operating on a very lean budget and supporting a very diverse global wildlife research community.

Ethical aspect: Please detail the way the solution addresses any of the main ethical aspects, including trustworthiness, bias, gender issues, etc.

Wild Me provides free access to Wildbook instances (installed versions of Wildbook open source software and AI models configured for specific species and directly supported by Wild Me engineers) to a globally distributed wildlife research community. We strongly believe that local efforts provide the most conservation impact, and bringing very technical AI capabilities to local field biologists (regardless of gender, race, or nationality) who do not have access to the needed technical skills, funding, or AI tools is at the core of our mission. We also provide our technology as open source, ensuring independent efforts can proceed without Wild Me and even advance the state-of-the-art in AI and human-computer interaction (HCI). Wild Me builds Wildbook with internationalization and localization in mind, developing the software to support multilingual users across the globe, and our current userbase extends across six continents for marine and terrestrial species. However, we recognize that wildlife data can be locally sensitive (e.g., needing protection from poaching efforts), and therefore we build Wildbook with information security in mind, providing the foundation for software-enabled collaboration across borders and projects but providing “silo security” that can protect the data of individual researchers and animals from malicious review. Overall, Wildbook seeks to support wildlife research and conservation efforts without bias and ensure that we are a part of locally-driven, ethical, and highly successful efforts to combat the Sixth Mass Extinction and support SDGs 14 and 15 with powerful and equitable AI.


International Research Centre in Artificial Intelligence
under the auspices of UNESCO (IRCAI)

Jožef Stefan Institute
Jamova cesta 39
SI-1000 Ljubljana




The designations employed and the presentation of material throughout this website do not imply the expression of any opinion whatsoever on the part of UNESCO concerning the legal status of any country, territory, city or area of its authorities, or concerning the delimitation of its frontiers or boundaries.

Design by Ana Fabjan