First-of-its-kind multispecies AI model to detect illegal wildlife trafficking is ready to roll out to airports
The first-of-its-kind multispecies artificial intelligence model to combat the $23 billion illegal wildlife trafficking industry has been developed by Microsoft.
Project SEEKER can be easily installed in luggage and cargo scanners at airports, ports, and borders, and will automatically alert enforcement agencies when it detects an illegal wildlife item. Officials can then seize the objects, which can be used as evidence in criminal proceedings against the smugglers.
Project SEEKER recorded a more than 70%
successful detection rate
Microsoft has trialled the Azure-based technology at Heathrow Airport, where it scanned up to 250,000 bags a day. It recorded a more than 70% successful detection rate and is particularly effective at identifying ivory items such as tusks and horns.
Clare Barclay, Chief Executive Officer at Microsoft UK, said: “The untapped potential of AI and machine learning can help solve some of the world’s most complex environmental challenges. Our first-of-its-kind multispecies AI model Project SEEKER can help tackle the wildlife trafficking trade, while protecting animal ecosystems. The importance of collaboration and partnership with more organisations couldn’t be greater as we look to protect the environment and the world’s most endangered species.”
Illegal wildlife trafficking is among the five most lucrative global crimes in the world, impacting more than 7,000 species of animals and plants. Project SEEKER’s AI model has been built to stop illegal wildlife trafficking in its tracks and is optimised to detect illegal wildlife products trafficked in luggage and cargo.
There is little data on illegal wildlife trafficking – a problem Project SEEKER is looking to solve – but in 2019 the UK Government revealed that in a single month Border Force CITES team officers in this country made 168 seizures containing thousands of illegal wildlife products.
Illegal wildlife trafficking rates fell sharply during the early months of the COVID-19 pandemic as countries closed their borders but there are fears this criminal activity is growing again. Pangolins, rhinos, elephants, and tigers are believed to be some of the most at-risk animals for their body parts but many other species from across the world are also targeted.
The illegal wildlife trade often goes hand-in-hand with other organised crimes, including corruption and money laundering.
Project SEEKER, which is a Microsoft AI for Good research project, has been trained to identify animals or products such as bears, polar bears, leopards, lions, and illegal products used in medicines. The Royal Foundation, the Duke and Duchess of Cambridge’s philanthropic organisation which runs a taskforce dedicated to tackling the illegal wildlife trade, believes technology can be a crucial tool for global enforcement and collective private and public sector action.
Developers have trained Project SEEKER by uploading images of animal parts and manually labelling them, so the AI model can spot them automatically in the future (note: this video does not have sound)
The Duke of Cambridge visited Microsoft’s UK headquarters to hear about the potential of this technology as part of his work with The Royal Foundation’s United for Wildlife programme. To support the development of this new technology, the Project SEEKER team was able to benefit from United for Wildlife’s global network of expertise on the illegal wildlife trade. In addition, United for Wildlife will be working with its partner organisations in the transport sector to support the global roll out of the SEEKER capability.
United for Wildlife aims to make it impossible for traffickers to transport, finance or profit from illegal wildlife products by building crucial relationships between the transport and finance sectors, Non-governmental organisations and law enforcement agencies and encouraging the sharing of information and best practice between these stakeholders. United for Wildlife has been working with organisations like Microsoft to raise awareness of technology that can support efforts to disrupt the criminal trade of wildlife products globally.
HRH The Duke of Cambridge met Microsoft UK Chief Executive Clare Barclay during a visit to Microsoft’s UK headquarters in Reading on Thursday. His Royal Highness was there to learn more about Project SEEKER as part of his work with The Royal Foundation’s United for Wildlife. HRH The Duke of Cambridge and Clare Barclay took part in a Microsoft Teams call with ranger teams in Africa to learn more about the issue of illegal wildlife trafficking, before meeting UK Border Force staff, the team behind Project SEEKER and Microsoft apprentices
Speaking ahead of the visit, Lord William Hague, Chairman of the United for Wildlife Taskforce, said: “The illegal wildlife trade is among the five most lucrative global crimes and is often run by highly organised criminal networks who exploit our transport and financial systems to move illegal animal products and their criminal profits around the world.
“This is a hugely complex global issue, but when institutions including transport, technology, financial services and enforcement agencies work collaboratively to share knowledge, expertise, and information it enhances our ability to detect and dismantle the sophisticated criminal networks that sit behind each trafficking act. Working in partnership with the public and private sector is crucial if we are to stop this illegal trade for good.”
Project SEEKER has been developed in partnership with research trial partners, UK Border Force CITES team, Heathrow Airport and UK security and screening technology vendor Smiths Detection.
Jonathan Coen, Director of Security at Heathrow Airport, said: “Project SEEKER and our partnership with Microsoft and Smiths Detection will keep us one step ahead of traffickers, by exploring new technology that will help us protect the world’s most precious wildlife.”
Developers have trained the Project SEEKER algorithm by uploading images taken of animals and their body parts. The system then takes a 3D scan of a bag, generated by Smiths Detection’s X-ray scanners, rotates the image to get a complete view of the luggage, filters out unrelated objects such as clothing and isolates any object that matches the pictures it was trained on.
The algorithm can be trained on any species using a limited dataset in just two months. As it’s still in the trial phase, Microsoft is calling on conservation organisations, law enforcement agencies and major transport hubs to deploy Project SEEKER and share data of rarer subspecies to improve the AI model’s capabilities.
Successful seizures by authorities have the potential to provide important, shareable insights into the illegal wildlife trafficking supply chain, including where the items originated, where they moved to and their destinations. This also sheds light on the identity and methods of traffickers.
Because Project SEEKER has been built on Microsoft’s global Azure cloud platform, that data can be analysed to build a comprehensive and detailed picture of illegal wildlife trafficking across the world.
Daniel Haines, AI Specialist and Project SEEKER Lead at Microsoft, said: “Illegal wildlife trafficking has a devastating effect for the decline of species and earth’s natural environments. It is a complex illicit trade but with the right AI intervention deployed in the right places, we have a real possibility of dismantling it. Project SEEKER shows the potential for data and AI to enable enforcement teams to crack down on wildlife trafficking like never before.
“Improved rates of detecting illegal wildlife trafficking at transit hot spots is just the start. The data captured by authorities will allow them to create a clear picture of where smuggling starts, its routes and destinations, leading to a more effective and collaborative approach to stamping out these criminal networks.”
Azure and its storage, machine learning, IoT and DevOps tools are compatible with many other systems, so airports usually don’t have to modify their existing digital security operations or scanners. Those scanners are already excellent at spotting firearms, explosives and drugs, but Project SEEKER can be trained to help authorities spot more of these items, too, and tackle criminal networks worth hundreds of billions of dollars.
Project SEEKER is the latest biodiversity initiative involving Microsoft. The company has helped to launch WildMe, a citizen science effort to build a digital library of all endangered species to help with conservation, while in 2020 Microsoft announced it aims to use data and digital technology to enhance environmental decision-making.
To learn more about Microsoft’s sustainability and biodiversity commitments, click here.