A button is an invitation only if its intended user can operate it. For a giraffe, a touchscreen at human height would be a poor starting point. An object within reach of a tongue or a sensor beside the head offers a different possibility: let the animal make something happen, then watch what it does with that opportunity.
Alana Grant, Vilma Kankaanpää and Ilyena Hirskyj-Douglas explored that possibility with five female giraffes at Blair Drummond Safari Park. Their question was partly about sound and partly about access. Could giraffes use an interface that played audio, and would the method of operating it change their engagement?
That distinction matters. A device can register many interactions because it is easy to encounter, because it attracts investigation or because an animal wants the sound. Counting activations does not automatically separate those explanations.
Building around a giraffe's movements
The team worked with keepers and familiar objects. Their touch-based prototype placed equipment inside a hanging object adapted from the kind of feeding toy the giraffes already encountered. Movement from a giraffe's nudge triggered playback. The device did not dispense food.
The second prototype used a proximity sensor outside the enclosure barrier. Bringing a head into its detection area could start the sound without touching the equipment. Photographs of the installations show why these were different invitations: one required manipulating a suspended object, while the other could respond to an animal approaching a particular place.
The touch device was studied first, over 30 days; the proximity device followed over 24 days. Each had an initial silent period, a period with audio and a final return to silence. During the audio phase, the available sound changed between recorded giraffe hums and white noise according to the study schedule. Giraffes were not choosing simultaneously between two labelled sound stations.
Device records were checked against camera footage to remove activations caused by people or irrelevant events. The researchers also took account of operating time, because the systems were not available for equal numbers of hours. Most interaction occurred when the animals were indoors at night.
These design decisions made the study more informative than simply putting a speaker in an enclosure. They did not remove every difference between the interfaces. Position, sequence, novelty and the movements required all varied alongside the trigger mechanism.
A more accessible trigger is not a favourite sound
The researchers reported more frequent interaction with the proximity interface, while interactions with the touch device lasted longer on average. This suggests that the physical arrangement affected use. It does not establish a general preference for one interface independently of its location and position in the study sequence.
A giraffe passing a sensor can activate it with less directed manipulation than one pushing a hanging object. The proximity device's position also meant that activities around that part of the enclosure could bring an animal into range. Camera verification helps identify a giraffe as the trigger; it cannot by itself show that triggering was the animal's purpose.
Neither interface produced a clear statistical preference between humming and white noise. That is narrower than saying the sounds were equally attractive, or that the giraffes disliked both. The experiment's data did not distinguish them convincingly.
Interest also changed over time. Use during the final silent period was lower than during the initial silent period. The authors considered whether sound might have discouraged engagement, but declining novelty offers another explanation: an unfamiliar object can be interesting before any audio is added. Figure 6 places interactions along the study days and makes that history visible. Initial silence and silence after several weeks are not identical experiences.
The study also contains inconsistencies between some reported operating totals and tabulated counts. Those prevent the precise interaction totals from carrying the interpretation. The more durable result is the comparison the experiment opens up: access, duration of engagement, novelty and sound selection are separate features of an animal-controlled system.
Why a giraffe voice needs its own evidence
Using recorded hums might seem an obvious route to meaningful enrichment. But the original humming study by Anton Baotic, Florian Sicks and Angela S. Stoeger established something more limited: giraffes in zoos produced sustained, low-frequency vocalizations during overnight recording.
Those researchers could describe the sounds' structure but could not identify the individual callers or determine their function. They did not demonstrate that hearing such recordings is pleasant, familiar or beneficial to another giraffe. A genuine animal sound is therefore a useful experimental material, not an automatic welfare intervention.
In the interface study, keepers reported no apparent adverse reactions or changes in the giraffes' established repetitive behaviours. That is relevant monitoring, but it was not a blinded, controlled demonstration of improved welfare. Enjoyment, control and beneficial effects would each need evidence beyond activation logs.
The work provides an initial way to ask better questions. Reversing the order of devices, keeping locations comparable and offering well-matched simultaneous options could help separate the interface from the sound. Observing whether an animal deliberately returns after hearing a particular output could add meaning to its next activation. The achievement is that giraffes were given an operable way to influence their surroundings; understanding what they wanted from it remains the next investigation.
Sources
- Grant, Alana; Kankaanpää, Vilma; and Hirskyj-Douglas, Ilyena (2023). “Hum-ble Beginnings: Developing Touch- and Proximity-Input-Based Interfaces for Zoo-Housed Giraffes’ Audio Enrichment.” Proceedings of the ACM on Human-Computer Interaction, 7, ISS, Article 434. Open author version.
- Baotic, Anton; Sicks, Florian; and Stoeger, Angela S. (2015). “Nocturnal ‘humming’ vocalizations: adding a piece to the puzzle of giraffe vocal communication.” BMC Research Notes, 8, 425. Related recording study.
