A man like Bruce Lee is made in his mannerisms.

There is, of course, the matter of appearance — the shaggy mane and sharp mandible, the taut musculature and furrowed brows. But these details are almost extraneous, mere accessories to adorn his movements. The true marvel of Bruce Lee is his gestural charisma: the snake-charmer swagger that, when directed against his enemies, distracts from the snakelike speed of each punch and kick. He wields his strength like a switchblade, controlled and deadly. Devoted fans are familiar with this fatal dance between hero and villain. Lee, always the hero, will inevitably triumph against all odds.

Lee died in 1973, on the brink of international stardom. At thirty-two, he didn’t live long enough to triumph in Hollywood, but death immortalized him as a Hollywood legend. When Lee died, he had about thirty acting credits to his name, having started early as a child actor. Only four, however, were for movies in which he was the leading man, flexing as the martial arts genius that he is posthumously recognized as: The Big Boss (1971), Fists of Fury (1972), Way of the Dragon (1972), and Enter the Dragon (1973). The films arrived in quick succession, one after another, as studio executives sought to capitalize on Lee’s insurgent popularity. Still, they were low-budget, schlocky flicks with few noteworthy qualities. The fight sequences easily blur together, the dialogue is poorly dubbed, and the villains and the storylines are cut from the same archetypal cloth. I don’t remember the names of any of Lee’s characters, nor the particularities of the plots. All I care to remember is Lee—almost always shirtless and always kicking ass.

One watches Bruce Lee to witness the miracle of his mortal body. His deft command over every movement transforms flesh into an instrument of pure expression. For most on-screen actors, the face is the primary vehicle to articulate emotion; the camera lingers on it to capture the feelings it invokes. Lee did not privilege the face, though it was no less important1 than his limbs in the kinesics of his acting. Yet his face is what we covet, for it confirms what his body’s fluid magnetism already implies: He is not just any Asian actor, any martial arts fighter. He is Bruce Lee.

**

In 2019 two filmmakers acquired the rights to use James Dean’s image to cast the deceased actor, who died in 1955, in a Vietnam War–era action drama called Finding Jack. Dean’s performance was to be constructed via full-body CGI, based on actual footage and photos. Still, this would require a body double (if not multiple) on set, and another actor would voice Dean’s character. The project has since been canceled, as the notion of Dean’s on-screen resurrection disturbed actors and viewers alike. This does little to change the fact that existing special effects technology can already convincingly replicate actors, dead or alive.

When Robert Downey Jr. was unable to film the final scenes of Iron Man 3 (2013), Marvel superimposed the actor’s face onto a body double. The CGI was deployed with Downey’s permission, but most digital re-creations done posthumously tend to skirt various ethical lines. For one, a dead person can’t grant permission for filmmakers to use their image and likeness. Context matters too, depending on how the technology is employed: Is a person’s image being re-created for public viewing and profit or for their loved ones in a private setting?

A 2021 documentary on Anthony Bourdain, for example, was controversial for using AI-generated audio of its subject’s voice without explicit disclosure to audiences.2 Meanwhile, a hologram of Robert Kardashian, commissioned by Kanye West for Kim Kardashian’s fortieth birthday, is arguably more uncanny in its effects but doesn’t prompt the same ethical questions. Regardless, with the proliferation and advancement of deepfake technology, these digital revivals have become much more common, and such hollow renderings are considered an homage to the person’s memory, a symbol of remembrance for families and fans alike.

For me, holograms like Robert Kardashian’s induce a dizzying sense of dread. Watching the video posted by Kim Kardashian on her Twitter3 feels like blasphemy: the deceased person’s essence is reduced to that of a puppet, programmed to perform certain identifiable mannerisms and utter familiar phrases. No matter how accurate the hologram appears to be, the intent to “revive” the deceased strikes me as necromaniacal. It’s distinctly uncanny, in the Freudian sense, blurring the boundary between fantasy and reality. The viewer is thus confronted with “the reality of something that we have until now considered imaginary, when a symbol takes on the full function and significance of what it symbolizes.”4

In his essay on the uncanny, Freud posits that “an element of intellectual uncertainty” is essential in linking the uncanny to death, which is what we truly fear.5 However, the most frightening aspect of these undead holograms is their lack of uncertainty. We know the person is dead, yet we accept—even delight in—their visual reincarnation. They are dead until they are technologically reanimated at the living’s beck and call.

**

Lee is a favorite among amateur deepfake creators who delight in experimenting with the technology. There are deepfakes imagining Lee as part of the cast in films like Shaolin Soccer, The Matrix, The Terminator, Dragon: A Bruce Lee Story, Ip Man 3 and 4, and Once Upon a Time in Hollywood. Unlike the holographic reanimations, these facial substitutions are not quite as eerie; the tampered footage is not convincing, even to my untrained eye. Perhaps the point is not to convince the viewer (for that would make the videos uncanny) but to fulfill a fictional reality where Lee became a Hollywood star. The reanimations serve as a visual aid for us to imagine Lee in a multiplicity of roles. A deepfake that’s clearly fictional, in this sense, feels safe. Rather than uncanny, it conveys an image that is “much richer than what we know . . . it embraces the whole of this and something else besides, something that is wanting in real life.”6

One of my favorite Bruce Lee deepfakes casts him as Tom in the 2009 romantic comedy 500 Days of Summer.7 Tom is so unlike most of the roles that Lee has been typecast in, in both life and death: a pining romantic, a man who’s given up on his dreams of being an architect. It was this contrast between Tom’s character and Lee’s on-screen persona that elicited my unexpected tenderness toward Lee’s Tom, when he faces Summer on the bench, his hair slicked back. In that scene, Lee’s features settle so comfortably over Joseph Gordon-Levitt’s that I was willing to suspend my disbelief, however briefly. To quote James Baldwin, I wanted to “surrender to the corroboration of [my] fantasies as they are thrown back from the screen.”8

I’ve found these fantasies to be less tantalizing the more “real” they purport to be, especially with deepfakes of martial arts biopics like Ip Man or Dragon, in which Bruce Lee is a character in the film, played by a living actor. With these edits, Lee’s face is seamlessly blended over the actors who were cast to embody him. Still, something isn’t right. The actor’s build is too wide, the high-definition camerawork feels anachronistic, and the fight sequences reveal the simulacrum’s ultimate insufficiency. There are, of course, distinctions worth making between these deepfakes: a deepfake that replaces Joseph Gordon-Levitt with Bruce Lee indulges in a fulgent vision of Lee as a movie star, compared to the more realist substitutions done in Bruce Lee biopics, where Lee’s face is pasted atop the actor playing his on-screen self. Yet even the “best” deepfakes, the ones that try to cast Lee as himself, are only simulating a shallow trace of the real Lee.

In his book Speech and Phenomena, the philosopher Jacques Derrida describes trace as an absence of presence: “The trace is not a presence but is rather the simulacrum of a presence that dislocates, displaces, and refers beyond itself.” In Derrida’s usage, trace is a semiotic concept (related to one’s understanding and interpretation of signs), but I find it to be a useful term in analyzing Lee’s cinematic afterlife. In these deepfakes, Lee’s image is “simultaneously traced and effaced, simultaneously alive and dead.” He is alive, but superficially so: His being is “[simulated] in its preserved inscription”9 at the expense—and the effacement—of another actor’s features.

**

I spent my summer watching second-rate martial arts movies, searching for traces of Lee. In the wake of Lee’s death, filmmakers across Asia began casting Lee lookalikes in derivative martial arts movies. These films were done in poor taste to exploit his international popularity, and none of the profits went to Lee’s family.

Studios even had actors go by similar on-screen names, like Bruce Li, Bruce Le, Bruce Liang, and Dragon Lee, effectively pigeonholing any promising Asian actors into the Bruce Lee typecast. This gave rise to a subgenre of martial arts movies dubbed “Bruceploitation.” Profit is an essential byproduct of any exploitation flick, and Golden Harvest, the production company that backed Lee’s final four films, was keen to eventually supplant him with another Asian martial arts star. His talent, so these executives thought, was fungible. But first, they wanted to milk the Lee craze for all it was worth.

Most Bruceploitation movies are not very good and only worth watching today for these bizarre traces of Lee. But for audiences in the 1970s, these copycat films temporarily filled the void that he left behind. Each Bruceploitation actor had a distinct approach to conveying his mannerisms. (Ardent fans of the genre believed that the films squashed the star potential of promising actors like Ho Chung Tao because filmmakers were intent on casting them in Lee’s posthumous shadow.) Still, the dramatic limitations of action cinema inhibited their performances. Unlike later biopics, like Dragon, that sought to be a homage to Lee by honoring his off-screen complexity, these cursory performances are more akin to a Saturday Night Live impression.

The Game of Death, released in 1978, is the most famous of the Bruceploitation flicks, for it contains footage of Lee. Lee had died while working on the film, which he wrote and directed, so Golden Harvest enlisted Enter the Dragon director Robert Clouse to rewrite the script and complete Game of Death alongside Raymond Chow and other Lee collaborators. While over one hundred minutes had already been shot, Clouse used only about twelve minutes of actual Lee footage in the final cut. The rest of the film was framed around two Lee stand-ins, Kim Tae Chung and Yuen Biao, who look nothing like Lee, but whose faces mostly appear indirectly. In fact, Clouse had pieced together an entirely new film from the original’s scraps (and scraps of other Lee films, including Way of the Dragon and Enter the Dragon).

The result is a bizarre and baffling postmodern composition that, while crass in its commercialized manipulation of Lee’s image (even in death), is wholly dependent on the death of its star, both on-screen and off. Clouse’s Game of Death is a mystifying collage of a post mortem cash grab, oscillating between fiction and reality, re-creation and imitation. It becomes, as novelist Tony Tulathimutte describes it, “a bizarre sequence of layered simulacra, of double entendres and trifocal ironies, overspilling the boundaries of its own fictional ontology onto reality.”10

The film follows Billy Lo, a Lee-like martial arts movie star, who is pursued by gangsters. While filming on set one day, Billy is shot in the face. He survives, but his damaged face has to be reconstructed via plastic surgery. Billy takes this opportunity to fake his own death and hunt down the gangsters who tried to kill him.

Clouse not only frames the movie’s plot around Lee’s death, he borrows footage from Lee’s real funeral for the fictional Billy’s funeral. In reworking this plot, Clouse would’ve been able to substitute Lee for a different actor or keep Billy’s face bandaged altogether. However, The Game of Death opted to abandon all notions of continuity and logic. While Lee’s stand-ins play Billy for most of the film, their faces are concealed or obscured so that random close-ups of Lee’s face can be stitched in throughout. Still, every clip of the real Lee feels asynchronous. In one early scene in Billy’s dressing room, Lee’s frozen face is haphazardly pasted onto his body double, and the result is as unbelievable as it is comical. The shots of Lee don’t match up to the rest of the movie, and his facial consistency is at odds with the plot: Why have Billy get plastic surgery in the first place?

In its blatant attempt to placate the audience by giving them Lee, Clouse’s Game of Death undermines itself. The film betrays its inherent commercialism by sacrificing any attempt at visual or narrative coherence. It reveals itself to be a parasitic creation, like a twin that absorbs its fetal sibling in the womb, emerging victorious. But at no point does the audience mistake the real for the fake. The specter/spectacle of Lee haunts Game of Death. His presence overwhelms the film’s climax and final minutes. I’ve interpreted Clouse’s decision to close Game of Death with Lee’s original footage as an act of resignation, though it may very well be oblation. Regardless, he did so with the understanding that audiences would not sit through an entire film once Lee’s twelve minutes were up. It was Clouse’s trump card in a movie full of tricks and shoddy sleights of hand.

But at no point does the audience wonder: What’s real and what’s fake? We can sense the real Lee and his command over the camera. The rest are just cheap clones, mass-produced action toys made in his image. There may be a few rare instances where you can glimpse a hint of him in a grimace or a cocky nod. Like the 500 Days of Summer deepfake, the illusion is quickly shattered. What lingers is the trace of him, felt in the loss: Bruce Lee’s utter absence becomes fully realized in his imagined presence.