Is artificial intelligence the solution beyond the fence?

Feral Herald |

It’s not every day you can celebrate native species being removed from the threatened species list! But earlier this year, experts have concluded that 29 have recovered enough to be taken off, with native mammals making up up 15 of those species. Predation from feral cats and foxes was the key threat to most of the recovering mammals.

Their recovery is largely thanks to emergency interventions in the form of a relatively small network of islands and fenced havens where feral cats and foxes have been excluded, eradicated or at least substantially reduced in number.

While this strategy has proven invaluable to stave off feral cat and fox-driven extinctions and declines, it is far from a permanent solution if our aim is to restore thriving interconnected landscapes.

The greater bilby (left) is one of the 29 species in recovery and no longer meet the criteria for being listed as a nationally threatened species. But it’s smaller, feisty relative the Yallara (lesser bilby; right) is one of over 30 native animal extinctions that feral cats and foxes have already played a role in since colonisation. Photo by Bernard Dupont and illustration via the British Museum (Natural History).

Beyond fenced havens

The larger challenge remains the sea of destruction wrought by feral cats and foxes beyond fenced havens.

The greater bilby and the burrowing bettong are two of the 29 native animals that have recovered well enough to come off Australia’s threatened species list. But both species used to inhabit over half the mainland. So despite being pulled back from the brink of extinction, they are still missing from most of their native ranges, along with the ecosystem services they provide. This is hardly a recovery.

A 2019 study shows us how important havens currently are for native mammals that are particularly susceptible to feral cat predation. Tracks of spinifex hopping mice are much higher inside the Arid Recovery Reserve, then drop off sharply outside the fence. Figure from Moseby et al. 2019. Photo by Nathan Johnson via Flickr, CC BY-NC-SA 2.0.

Using AI to protect native species from cats and foxes

One of the most promising new tools in development that is helping to protect native wildlife outside predator-free havens is the Felixer.

These are solar-powered grooming traps that target cats and foxes with a small toxic gel bait. It has smart technology that automatically avoids firing on humans and other non-target species. The bait is later ingested by targets, particularly cats since they are habitual groomers and lick off the gel.

Thylation, the group behind Felixers, has now integrated camera-based artificial intelligence into newer models, on top of the existing LiDaR rangefinder sensors, to improve their target detection. Recent models have the functionality to avoid firing on any animal if they detect a bluetooth pet collar nearby. By affixing one of these bluetooth collars to pet cats and dogs, they can be safeguarded from being targeted.

Felixers have been developed to bait cats while avoiding non-target species. The devices use inbuilt rangefinder sensors and, in newer models, camera-based artificial intelligence to make detections. Photo by the Thylation Foundation.

Felixers have been deployed for research and testing purposes in numerous remote areas, and are quickly gaining attention as a new conservation tool for conducting highly targeted feral cat control with minimal non-target impacts. In February 2023, Felixers were registered for use across Australia, subject to state controls.

Do Felixers work?

Thanks to a very generous donation we have been able to partner with the Thylation Foundation to provide Felixers into critical habitats to support threatened species conservation and rewilding programs. One of those projects has been the recent deployment of Felixers to Arid Recovery, a 123 km2 fenced conservation area in northern South Australia. The reserve is home to native species like kowari and bilbies that are highly susceptible to feral cats and foxes.

In their first month of operation, 11 feral cats were verified as having been detected and 3 of those were targeted by a Felixer with a gel bait. The remaining 8 feral cats were not targeted because the device either wasn’t confident enough in its detection or wasn’t lined up to land the gel bait on the cat. While this means only 27% of detected cats were targeted, the Thylation Foundation says it expects an upcoming algorithm update will increase the success rate of firing the bait on detected cats to around 70%.

During January 2023, 8 greater bilbies (left), 8 burrowing bettongs (centre) and 5 re-introduced chuditch (western quolls; right) were also detected by the Felixers in and around the Arid Recovery Reserve but none were fired upon. These species are still missing from enormous expanses of their natural habitat that lies outside of introduced predator-free islands and fenced sanctuaries.

We can’t be sure exactly how many of the three feral cats that were successfully targeted then went on to ingest the gel bait and be removed from the ecosystem. And it will be a waiting game to see what longer-term impacts these feral cat removals have on the native wildlife in the area. But previous efforts can give us some insight into what we might expect.

An earlier trial of Felixers at Arid Recovery successfully targeted 33 cats, with none of the cats that could be individually identified being detected again during the study period. By the end of the six week trial, the detection rate of feral cats had reduced by nearly two-thirds. None of the 1,024 non-targets like bilbies, birds, quolls, bettongs, lizards, kangaroos and humans were targeted by the devices.

Observations by ecologists on the ground suggest threatened kowari are already experiencing higher survivorship within the fenced reserve. Since the juvenile kowaris are small enough to disperse out through the fence where Felixers are now removing feral cats, we may begin to see increased survivorship outside the fenced reserve as well.

A wildlife camera captures young kowaris emerging from their den at the Arid Recovery reserve.

Elsewhere in Australia, early results show Felixers are already helping drive recoveries of native birds and giant geckos on Christmas Island and golden bandicoots in north-western NSW.

Adding Felixers to the toolbox

At a feral cat symposium in February 2023, Invasive Species Council campaigners heard how one of the major shortcomings of feral cat management across Australia is bureaucratic complexity. A confusing mess of legislation and protocols that differs between states, territories and even local governments means that most land managers are being prevented from being able to use tools that should be in their feral cat and fox control toolbox.

The NSW government bans the use of the PAPP-based bait Curiosity outside of National Parks and Wildlife Service lands. And the Victorian government has issued a blanket ban on the use of 1080 baiting to control feral cats, one of Australia’s most important tools for professional land managers. This, together with barriers preventing the testing of PAPP-based Felixers, means Victoria is also the only state or territory that prohibits the use of Felixer traps, even under research permits for conservation purposes.

While it is essential to ensure that all invasive species management is safe, humane and effective, it’s our native wildlife that will ultimately suffer the consequences if governments over-complicate approval processes and demand unreasonable conditions for tools to control feral cats and foxes. Tools like Felixer traps and the artificial intelligence plugged into them will only improve, and new tools will emerge. We can’t fall into the trap of letting regulations drag behind technological jumps in humane and effective feral cat control that can offer our native wildlife new opportunities to survive and recover.

To extend the recent recoveries we’ve seen in native animals like greater bilbies, burrowing bettongs and chuditches beyond introduced predator-free havens, we need to give land managers every effective tool that is available.

The recent conservation success stories on islands and within sanctuaries cannot allow us to ignore the deadly reality that awaits our native wildlife outside these havens. By combining trapping programs, landscape-level baiting and well-planned Felixer deployments, we can humanely and effectively release Australian wildlife from the pressures of feral cats and foxes that have already sent too many native species to an early grave.

Share this post


Other posts

Email Updates

Sign up for our ebulletin the the Feral Herald and regular campaign updates.

Recent Stories


Help protect NSW!

Our expert team has written a list of policy asks detailing exactly what the next NSW government needs to do to stamp out some of the worst invasive species impacts across the state. But they will only become a reality if every key political candidate at the 2023 NSW state election hears about it from you!

Dear National Deer Management Coordinator,

Please accept this as a submission to the National Feral Deer Action Plan.

[Your personalised message will appear here] 

I am very concerned about the spread of deer and am pleased that a national plan has finally been developed. Without urgent action, funding and commitment from all levels of government it is clear that feral deer will continue to spread and damage our environment.

The feral deer population in Australia is growing rapidly and spreading across the country, damaging our natural environment, causing havoc for farmers and foresters and threatening public safety. Unlike much of the world where deer are native, our plants and wildlife haven’t evolved to deal with these heavy hard hooved animals with a voracious appetite.
With no natural predators and an ability to adapt to almost all environments, they could occupy almost all of Australia unless stopped. Despite this, state and territory governments have been slow to respond and in Victoria and Tasmania they are still protected by law for the enjoyment of hunters.

This plan should be adopted by all governments but must also be underpinned by dedicated funding and clear responsibilities. A plan without funding or accountability is a plan that will fail and Australia cannot afford for this to fail.

In order to prevent the spread of feral deer and reduce their impact on our native wildlife, ecosystems and agriculture, I ask that the following recommendations be adopted for the final National Feral Deer Action Plan:

1. All federal, state and territory governments should adopt the National Feral Deer Action Plan and declare feral deer to be a priority pest animal species.

2. All federal, state and territory governments should commit to:

  • Contain deer to the existing large population areas.
  • Reduce and eradicate smaller and isolated populations.
  • Protect important environmental assets such as world and national heritage areas.
  • Develop and fund regional plans and strategies to manage deer populations which involve land managers across all tenures.

3. In order to drive action and the success of this plan, there should be dedicated Commonwealth funding and support for:

  • A permanent national feral deer coordinator position.
  • A permanent federal feral deer action committee with representatives from the commonwealth and state and territory governments and the environmental and agricultural sectors.
  • An ongoing public education campaign on feral deer.
  • A network of regional feral deer coordinators to drive local action across tenures.

4. The expected outcomes for the plan need to be more ambitious, with clear interim targets including:

  • Within one year, all States and Territories should have in place arrangements to implement the National Feral Deer Action Plan, including allocating dedicated funding for implementation.
  • Within one year, feral deer management plans should be developed for key environmental assets of national significance, including the Tasmanian Wilderness World Heritage Area, the Greater Blue Mountains, the Australian Alps, the Gondwana Rainforests and the Wet Tropics of Queensland.
  • Within five years coordinated landscape scale management should be in place where land owners, land managers, government and community are demonstrably working together.

5. A national feral deer containment map with three zones should be adopted. It should be more ambitious than the zone map in the current draft plan and there should be greater clarity in the naming of the zones. Improvements that should be adopted include:

  • Renaming the zones to better reflect the management intention to ‘Containment Zone 1’ (the current large population zone), ‘Containment Buffer Zone 2’ (the current buffer zone) and ‘Eradication and prevention Zone 3’ (the current small isolated population zone).
  • The NSW northern rivers area should be in the eradication and prevention zone as there are few feral deer currently in this region and eradicating isolated populations and preventing spread into this area is still possible.
  • The whole of South Australia should be in the eradication and prevention zone as eradication is the goal of the SA Government.
  • The Tasmanian region in the containment zone should be smaller to reflect greater ambition and potential for eradication of deer populations.
  • In eastern Victoria areas such as Wilson’s Promontory, Westernport islands and the Mornington Peninsula should be in the eradication and prevention zone.

6. There should be consistent laws and regulations across all states and territories that:

  • Recognise feral deer as a pest animal and treat them as such.
  • Establish a clear responsibility for all landholders and managers to be involved in feral deer control programs.
  • Set clear penalties to stop the wilful or negligent release of feral deer.
  • Prevent new deer farms in areas where no feral deer are present and phase out all deer farms in the eradication and prevention zone.
  • Enable enforcement of compliance, including on government land.

I support the follow principles being adopted in the final National Feral Deer Action Plan:

  • Feral deer are a pest and should be treated as such on all tenures, except on approved deer farms.
  • Federal, state and territory governments have a responsibility to fund the outcomes under this plan.
  • All land managers in areas where feral deer are present have a responsibility to be involved in feral deer control programs.
  • The focus of management efforts should be on eradication of isolated, satellite populations, protection of key environmental assets currently impacted and stopping the spread to new regions.
  • Feral deer control should be undertaken humanely, safely and professionally according to agreed protocols and all tools which meet this criteria should be adopted, including aerial control.
  • Funding for coordination, regional planning and community engagement is necessary for effective feral deer management.
  • Ongoing management and follow up control efforts are required to achieve long lasting results.
  • Rules and regulations should be consistent across jurisdictions and land tenures.
  • Recreational hunting is not an effective strategy for feral deer control and should not be relied upon.
[Your name]
[Your email address]
[Your suburb], [Your state]