Everyone Cried for This Robot. But It Was Never Dying
Sun Yuan and Peng Yu’s Can’t Help Myself became the internet’s favourite dying machine. The story was false. What the work actually reveals is more disturbing.
The internet gave this machine a life it never possessed and a death it never suffered. That may be the most revealing thing that ever happened to it.
In the videos, a black industrial arm bends over a pool of dark red liquid. It drags the fluid towards its base, turns, performs an awkward little dance and returns to the same impossible task. The floor is stained, the transparent walls resemble a cage, and after a few seconds steel acquires fatigue. We begin to want to rescue a factory instrument.
This is Can’t Help Myself, created in 2016 by the Beijing-based artists Sun Yuan and Peng Yu. It is also something the artists did not exactly make: a viral fable assembled from captions, edits, melancholy music and millions of projections. The distance between the artwork and that fable tells us more about contemporary visibility than another dutiful explanation of what the robot “means”.
The death that never happened
The story repeated across TikTok and Instagram was irresistible. The red substance was supposedly hydraulic fluid leaking from the robot’s body, while the machine had been programmed to scoop its own “blood” back towards itself. As it lost more fluid, it weakened. Its dances became desperate and eventually, exhausted by the work of keeping itself alive, it died.
There is only one inconvenience: this did not happen. The liquid was cellulose ether mixed with coloured water and was not necessary to keep the robot operating. The machine did not bleed to death. It was switched off.
The correction has sometimes been offered as if it settled the matter and sent everyone home slightly embarrassed. I think it does the opposite. The false account did not merely obscure the artwork; it created a second one, built collectively by people who needed to discover a dying creature inside a machine. The internet chose the version that could be felt instantly, even if it had to invent the wound.
This is where Can’t Help Myself becomes much more than a viral robot. It allows us to watch meaning being manufactured in public. One version was made from industrial hardware, sensors, software and coloured water. The other was made from repetition, identification and the small narrative lies that help images travel.
A machine built to notice transgression
Commissioned for the Guggenheim Museum’s exhibition Tales of Our Time, the installation combined a KUKA industrial robot, visual-recognition sensors, software, a rubber blade, red liquid and a transparent enclosure. Cameras detected when the fluid crossed a defined limit. The arm responded by dragging it back.
The machine alternated between containment and display. During pauses in the spill, it performed gestures that looked absurdly social: bowing, shaking, scratching and dancing for the people outside the glass. A cleaning mechanism became a performer because its movements resembled ours just enough. Apparently, anthropomorphism needs very little encouragement.
Yet the charming gestures should not distract from what the system is doing. It watches a boundary, detects a breach and corrects it. The Guggenheim’s account of the work connects the blood-like traces to surveillance and to the violence produced around guarded borders. The red pool is therefore less interesting as imaginary robot blood than as the residue of a system trying to keep movement under control.
The machine is neither an evil guard nor an innocent prisoner in any simple sense. It is an apparatus carrying out an instruction, which is closer to how power often operates now. Political decisions are translated into procedures, risk scores, databases, cameras and automated refusals. Once the decision has become a technical response, responsibility becomes difficult to locate. The machine has no hatred to confess and no conscience to interrogate.
This places the work in a productive conversation with Ai Weiwei’s political art. Ai often makes authority visible by exposing material evidence, erased histories or damaged bodies. Sun Yuan and Peng Yu show authority after it has lost its face and acquired a user manual.
Why the robot received our compassion
The popular interpretation was not simply foolish. Viewers recognised employment, burnout, depression, illness, care work and the daily maintenance of a self that never feels repaired. The robot’s repetitive labour genuinely invites those projections, and art does not remain the private property of its makers’ intentions once it enters other lives.
Still, our attachment to the machine deserves suspicion. A robot is an unusually convenient victim: innocent, photogenic and unable to complicate our sympathy. It does not arrive with documents that can be questioned, a history that can become politically inconvenient or demands that might cost us anything. We can mourn it generously because it will never ask to cross an actual border.
A migrant is rarely granted such uncomplicated compassion. The moment a person appears, empathy meets administration. Is the story credible? Is the danger sufficient? Is the body legal? Does this particular life deserve passage? The machine escapes the trial imposed on the human being, even though the original work points directly towards the technologies that conduct that trial.
This reversal is the uncomfortable core of the viral episode. Millions of people felt the robot’s imaginary suffering more readily than the real political violence condensed into the installation. They were not heartless. On the contrary, they felt a great deal. But feeling travelled more efficiently after consequence had been removed.
Contemporary culture has become extraordinarily skilled at displaying concern. A heart, a repost or a tearful comment records that an image reached us, and the platforms convert that reaction into another unit of circulation. Nothing in this exchange guarantees that the condition shown by the image has altered. We did not rescue the robot; we distributed evidence of our compassion for it.
When the algorithm becomes the curator
Museums still behave as if meaning can be stabilised by the correct quantity of wall text. They provide dates, intentions, materials and context, hoping the visitor will leave with an interpretation that has passed quality control. Can’t Help Myself demonstrates how little authority that explanation retains once an artwork reaches a platform designed around speed.
“A robotic system addresses surveillance and border control” is accurate but requires attention. “A robot is desperately trying to stop itself from bleeding to death” arrives already complete, with a body, a threat, a task and an ending. The algorithm did not have to censor the political interpretation. It simply rewarded the sadder story until the other meanings became harder to see.
This is how platforms increasingly edit culture. They rank interpretations according to their ability to produce immediate reaction, and repetition then begins to look like consensus. By the time the false story has appeared under a hundred videos, correcting it feels almost cruel. Fact interrupts a grief that people have already experienced as authentic.
The process also complicates the Western reception of Chinese contemporary art. Works from China are often filtered through familiar expectations of censorship, dissent, authoritarian control, spectacular cruelty and technological acceleration. Those histories matter, but they can harden into an export identity. The international success of Chinese Cynical Realism already demonstrated how quickly political ambiguity can be converted into recognisable imagery for an external market.
Sun Yuan and Peng Yu are less accommodating. Their installations solicit wonder, discomfort and sometimes disgust, but they rarely tell us exactly where moral innocence is standing. Looking itself becomes part of the machinery. This connects their work to a broader problem explored in feminist art across Asia, where visibility may produce recognition but may also expose, simplify or endanger the person made visible.
The problem with transparent walls
The enclosure around Can’t Help Myself is not incidental. Without it, we would see a robot operating in a room. Behind transparent panels, the mechanism becomes an exhibit, an animal in captivity and a threat that has supposedly been made safe. We can inspect everything while remaining comfortably outside.
Glass promises knowledge, but here it manufactures distance. The machine is completely exposed and no freer for being visible. Its task, its stains and its apparent exhaustion can all be seen without interrupting the programme. Perfect visibility becomes one of the conditions of confinement.
This is a mistake contemporary culture repeats with impressive confidence. We assume that bringing something into view changes its condition, although visibility often does little more than create an audience. The machine works because we are watching, and we continue watching because it works. Sympathy is folded back into the spectacle as fuel.
At the 2019 Venice Biennale, the installation entered one of contemporary art’s largest machines for producing attention. Biennials gather war, migration, ecological collapse, identity and bodily danger into an itinerary that can be consumed between photographs, coffee and the next pavilion. Can’t Help Myself did not escape that contradiction; it sharpened it. The visitors outside the glass were never neutral observers of a system elsewhere.
The standard reading says that we are the robot, endlessly working, gathering what we lose and performing cheerfulness during the pauses. That interpretation is understandable but becomes comforting when recognition is treated as an ending. “That robot is me” can function like therapeutic content: we feel seen by the condition and then return to it.
I keep returning instead to the people outside the enclosure. We are also the spectators fascinated by repetitive distress, the cameras detecting deviation, the institution framing the experience and the platform selecting its saddest fragment. We helped invent the robot’s death because a death made the image easier to love. The work’s sharpest achievement is not making a machine appear human; it is making human attention appear automated.
Who built the room?
The robot did not slowly bleed to death. That correction should not reassure us because the desire that produced the story remains intact. We wanted the machine to suffer in a legible way, and we wanted its labour to end in tragedy. Tears allowed the encounter to feel resolved.
But the installation resolves nothing. The liquid crosses the boundary, the robot pushes it back and the stain remains. The audience watches, records and circulates. Every component performs its role, including us.
This is why Can’t Help Myself remains important long after its first exhibition and its viral afterlife. It reveals how easily attention replaces action, how quickly exposure is mistaken for change and how compassion loses its politics when it demands no consequence. Millions recognised themselves in the machine’s captivity. Far fewer asked who had built the room.
Can’t Help Myself was commissioned by the Solomon R. Guggenheim Museum for Tales of Our Time and was subsequently presented at the 58th Venice Biennale in 2019. Further technical and exhibition information is available in the Guggenheim artwork record and the Venice Biennale archive.
