BRENNEN

Industries / Media, arts and recreation

Zoos and aquariums

Zoos and aquariums keep wild animals in human care so that people can see them, scientists can study them, and threatened species can be bred and, where possible, returned to the wild.

1828

The Zoological Society of London, set up in 1826 "to create a collection of animals for scientific study", opened its gardens in Regent's Park to its members; it opened them to the public in 1847 to help pay for the society.

londonmuseum.org.uk

32

Milestones achieved across 6 domains.

7

To watch for: announced, not yet confirmed outside.

11

Open deliverables. Nobody has shipped them.

In one minute

Where it started

The modern zoo began in London, where the Zoological Society opened a scientific collection to its members in 1828 and to the public in 1847. The same society opened the first public aquarium, the Fish House, in 1853, and the naturalist who stocked it, Philip Henry Gosse, coined the word "aquarium". In 1907 Carl Hagenbeck's park near Hamburg showed animals behind hidden moats instead of bars, the start of a long move toward exhibits that look like the wild.

Where it stands

A survey by the World Association of Zoos and Aquariums (WAZA) estimated more than 700 million visits a year and about US$350 million a year spent on wildlife conservation, both from 2008 figures. Species360, a non-profit, keeps the shared animal records of more than 1,400 member institutions in more than 100 countries. Zoo-bred animals have put species back in the wild: the scimitar-horned oryx left the "Extinct in the Wild" category in 2023, and California condors, all taken into zoos by April 1987, numbered 607 at the end of 2025.

What is open

Much of the checking is still done by people. Trained observers record behaviour by hand, and the published camera systems found here that spot pacing were each tested on two polar bears or two elephants. The records that would show what works are scattered: no public table pools the outcomes of elephant tuberculosis and herpesvirus cases, or of zoo-bred animals after release, and visitor learning is measured mostly by what visitors say they will do. Camera models, cheap water sensors and AI models that read papers and tag text are available today to close these gaps, none of which requires touching an animal.

Domains

AchievedWatch forOpen

Open in zoos and aquariums

3 of 11 open deliverables

Open · Animal care

A pacing detector for zoos

Software that reads footage from cameras a zoo already has and reports each day how long each animal spent pacing or swaying. Welfare scientists call these repeated, unvarying movements stereotypic behaviour and treat them as a warning sign. The detector would be trained and tested on clips from many zoos and species, and the labelled clips published.

Built with what exists

CCTV footage that zoos already record; Ultralytics YOLO11 to find animals in each frame; DeepLabCut, which Lund and colleagues used, for body posture; the Zürl team's public tracking code as a starting baseline; Label Studio for labelling clips.

Credit · unclaimed

Open · Veterinary care

Pool the elephant TB cases

A public, searchable table of every published case of TB in a zoo or circus elephant: species, age, results by test type, drugs and doses, side effects, and outcome, including what was found at necropsy (an animal post-mortem examination). It would be built by reading papers, case reports and conference proceedings, not by testing animals.

Built with what exists

Europe PMC's free literature search; Claude or a similar AI model to copy fields from each paper into a fixed form, with a veterinarian checking every row (see the Note "Seven more flagged"); Datasette, an open-source tool, to publish the table.

Credit · unclaimed

Open · Habitats and water systems

Early warning for tank water

A public set of life-support sensor logs from aquariums (temperature, pH, dissolved oxygen, oxidation-reduction potential, flow and pump power) with real faults marked by staff, and an open detector that flags faults before water quality drifts out of range.

Built with what exists

Logs that aquarium control systems already store; Python with scikit-learn's Isolation Forest and the PyOD library; the published MSCRED model that the thesis extended; a bench rig (a test tank with sensors and no animals) for creating faults on purpose.

Credit · unclaimed