What follows from it
A prerecorded parrot video and a responsive live parrot should not be assumed to be equivalent enrichment. Anything built for birds to watch should ask whether the bird can affect what happens next.
Bird Smart · mirrors and screens
A bird that understands a mirror is not necessarily a bird that recognises itself.
Those are separate experimental claims, and the gap between them is where almost every viral mirror story goes wrong. This page separates them carefully — and then turns to the one screen result with real consequences for anyone designing video for parrots.
The test itself
What the mark test is actually asking
Proven method
Gallup's 1970 procedure
After an animal has experience with a mirror, it receives a visually conspicuous mark on a body area it cannot see directly. Increased mark-directed examination or touching while the mirror is available is read as evidence that the animal relates the reflection to its own body.
Note how narrow that is. It is a test of visual mirror self-recognition, not a universal assay of self-awareness. An animal can fail because it is visually or motivationally unsuited to the procedure, and passing does not establish every form of human-like self-concept. Published positive evidence exists in several nonhuman groups — most prominently great apes, with positive reports also for dolphins and elephants — while other taxa remain debated. The honest framing is an evolving experimental literature, not a self-aware/not-self-aware checklist.
Gallup, Science 167, 86–87, 1970, 10.1126/science.167.3914.86Where parrots stand
Useful with mirrors, unproven at recognising themselves
- Proven
Greys can extract real information from a reflection
Pepperberg and colleagues tested African greys on mirror-mediated tasks in the 1990s. The birds could use mirrors to locate information or objects not directly visible. That matters cognitively — it shows a reflected scene can be treated as information about the external world. It does not establish that the bird identifies the reflected bird as itself.
- Proven negative result
Kea and Goffin's cockatoos did not pass
A 2019 study testing both species found no evidence of passing the mark test. Some behaviours suggested the birds distinguished different mirror configurations, and social responses changed with experience — but the critical mark-directed evidence never appeared.
van Buuren et al., Behaviour 156, 763–786, 2019, 10.1163/1568539X-00003514 - Proven negative result
African greys and hill mynas did not pass either
A 2021 comparative study found substantial attention to mirrors and changes in behaviours such as grooming, but all subjects failed the mark test.
So the defensible statement, as of August 2026, is that no parrot species has a robust, independently replicated demonstration of classic mirror self-recognition. That is a very different sentence from “parrots do not understand mirrors.”
Lin et al., Zoology 147:125942, 2021, 10.1016/j.zool.2021.125942 - Contested
The famous magpie result did not replicate
Prior, Schwarz and Güntürkün reported in 2008 that Eurasian magpies showed mark-directed behaviour toward coloured throat stickers viewed in a mirror, concluding this was mirror self-recognition and the first such result in a nonmammal. A close replication by Soler and colleagues, with a larger sample, did not reproduce the mark-test result. The replication authors were careful to note that a failure to replicate does not logically disprove the original — but the claim should be treated with caution.
So: magpies have a famous positive report that later failed a close replication. Not “magpies have proven self-awareness.”
Prior et al., PLOS Biology 6:e202, 2008, 10.1371/journal.pbio.0060202 · Soler et al., J. Comp. Psychol. 134, 363–371, 2020, 10.1037/com0000223
Five different claims
Keep the levels apart
A bird's behaviour toward a reflection can sit at any of these levels. Each is a separate experimental question, and none of them implies the next.
01
Social response
Responding to the reflection as though another bird is present.
02
Learned oddity
Learning that this particular “bird” behaves unusually and adjusting accordingly.
03
Spatial use
Using reflected information to find something not directly visible.
04
Contingency
Distinguishing live interaction that responds to you from imagery that does not.
05
Self-recognition
Relating the reflection to one's own body. Not yet demonstrated convincingly in any parrot.
Practical consequence: a bird that stops attacking a mirror, or uses one to find a treat, has not been shown to be self-aware. No product, page, or app should apply that label on that basis.
The result that changes screen design
Live mattered. Recorded did not, nearly as much.
Proven behavioural difference · preliminary interpretation
Nine parrots, six months, a clear preference
A 2024 experiment gave nine parrots a choice between live parrot video calls and recordings of previous calls. Over six months the birds initiated 65 live sessions versus 40 prerecorded, and spent roughly 561 minutes in live calls versus 142 minutes with playback. Mean engagement per live interaction was also longer.
The authors are appropriately cautious: a sample of nine prevents strong claims about what the birds conceptually understood. But the behavioural difference is robust, and it points at something specific — social contingency. The most interesting variable may not be photorealism at all. It may be reciprocity: I call, the other animal reacts.
That is an inference from an observed preference, not proof that parrots hold a human concept of live telecommunication. It is still the single most actionable finding on this page.
Hirskyj-Douglas et al., CHI 2024, 10.1145/3613904.3641938What does not follow from it
“Parrot video calling prevents loneliness” or “cures depression.” Caregiver reports of benefit are useful qualitative data, not blinded welfare measures.
Read alongside: the display-hardware side of this question lives on the senses page — ultraviolet colour an ordinary screen cannot reproduce, and flicker sensitivity that may reach the 90–100 Hz region under bright conditions. The interface and agency side lives in communication technology. For how these ideas relate to the applications in this house, see the companion guides.
Where the story gets stretched
Three mirror claims to retire
“She's having a conversation with herself”
A parrot talking at its reflection is not evidence of self-directed conversation.
Mirror use proves self-recognition
Using a mirror as a tool for finding things is a different, well-documented, and separate ability.
Failing means no self at all
Failing a sticker mark test does not prove the absence of every form of self-representation. It fails one specific visual test.
Continue
Where to read next
The wild social system these responses evolved in is covered in wild vocal culture. The machine-learning side of interspecies communication — including the 2023 study in which 18 parrots learned to initiate their own calls — is in communication technology. General experimental caveats are on the cognition page.
They see much more than a bird-shaped blob
Budgerigars discriminate conspecific identity information
In controlled same/different tasks, budgerigars have discriminated visual information from conspecific faces. Parrot social vision supports far more than a generic category of “bird.”
Brown & Dooling, Journal of Comparative Psychology, 1992The budgerigar “virtual demonstrator” playback study
A video-playback social-learning experiment in budgerigars is frequently cited as showing that birds can copy an action seen from a virtual conspecific. Our research pass located it only through a secondary PDF index rather than a clean publisher record. It is listed here with that provenance gap stated, and is deliberately not used to support any claim on this site until the primary bibliographic record is confirmed.