XU BING, stills from Dragonfly Eyes, 2017, video with color and sound: 81 min. Courtesy the artist.

Man, Machine or Dragonfly?

Also available in:  Chinese

The buzz surrounding Xu Bing’s new film Dragonfly Eyes (2017) is that it was created from cutting and compiling more than 10,000 hours of video surveillance footage and is possibly the first such “film” to combine the technologies of over 245 million global surveillance cameras and cloud computing. But the release of the 81-minute movie also revealed an interesting decision by Xu—one of China’s most esteemed contemporary artists—to introduce this work not through museums, biennials or galleries, but rather via the ostensibly more mass-market medium of cinema. 

Dragonfly Eyes premiered at the Locarno Film Festival in August of 2017 and was rolled out in other major film festivals—including those in Toronto and Busan—in the following months. The screenplay is by Chinese poet Zhai Yongming and writer Zhang Hanyi, and the story was pieced together by Matthieu Laclau, an award-winning film editor whose past collaborators include top directors in China, Hong Kong and Taiwan, such as Jia Zhangke and Sylvia Chang. Distribution rights are now being sold by a Beijing film agent to territorial markets all around the globe. As a project, it is very much a feature film. 

While festivals have eaten up Dragonfly Eyes on account of Xu’s star power, the film industry as a whole has not been especially kind. Entertainment media were quick to pan Xu as an artist unsuccessfully dabbling in cinema, with the Hollywood Reporter calling Xu’s film “a failed experiment” and Variety writing that “Dragonfly Eyes disappoints with its unsatisfying combination of an ultra-banal narrative told in a formally original way.” The film’s current rating on the review aggregating website Rotten Tomatoes is a mediocre 60 percent. On the other hand, Xu has at least established a beachhead. Even if the artist is on foreign ground, a whole new cultural sector is now discussing his work.

At the Busan International Film Festival, where I attended the Dragonfly Eyes Asian premiere last October, the crowd of actors, producers, distributors and agents—a group who judges a work, at least in part, on its box-office potential—directed blunt criticisms toward the film’s narrative. People described the boy-girl love story as clumsy, contrived, dull and even stupid.

They are, of course, correct. The plot is tawdry and fails at any sort of emotional engagement. But there is another level in which Xu succeeds, and that’s in presenting the deeply haunting strangeness of our new technological reality. Every moment is now be recorded, processed and evaluated; not only by humans, but also by smart and smarter computers and their omnipresent gaze. The film asks: In a world of global networks, who does the actual seeing? Man or machine? Its strength is in showing us the world from the perspective of a computer network, and this is what makes it so unnerving.

The film’s opening scene is silent, grainy, monochrome CCTV footage of a woman walking alone at night. Staring at her cell phone, she fails to see the edge of a bridge and stumbles into a canal. The video begins to fast-forward, and we see her struggle in the water and drown. This footage is as horrific as it is banal and dispassionate, for it is real. And it underpins everything that is to follow. 

The plot chronicles the journey of a young Buddhist nun—Qing Ting, or Dragonfly—who decides to leave her monastery and explore the wider world. She meets a young man, Ke Fan, who falls in love with her but is jailed after committing theft in order to gain her favor. While Ke Fan is in jail, Qing Ting disappears, and takes on a new online identity. 

One can imply that the drowning at the film’s start represents, both literally and symbolically, Qing Ting’s disappearance into the depths of her cell phone and in the scope of the film, as she departs from the real world for an alternate virtual life. The film’s title, Dragonfly Eyes, is also a metaphor for the all-seeing eyes of global, computer-controlled surveillance networks, which Qing Ting has, in a sense, merged into her new disembodied form.

The remainder of the film is taken up by Ke Fan’s search for Qing Ting, often narrated with lame voice acting and computer text-to-speak narration, belying the fascinating, discombobulating interplay between two characters with constantly shifting identities. We hear what sounds like a children’s radio play in the overdubbed audio, while the video shows us characters who literally change bodies from scene to scene due to the variety of surveillance footage used.

The production process involved a similar technological uncanniness, facilitated by recent advances in networked video. Xu originally attempted the project in 2013, intending to make the film without any cameras or camerapersons and only using found digital clips. However, he found he could only access a sufficient quantity of CCTV footage needed for the movie once China’s surveillance cameras linked into cloud computing in 2015, at which time he began production in earnest. His method of “filming” often relied on video search as a means to source clips for a pre-written script. For example, at one point during production, Xu recalled in an interview that “we needed footage of a car driving along a mountain road on a rainy night. First we checked weather forecasts to find out where it was due to rain. Then we locked on to a surveillance camera in that area. The next day, we checked to see if we’d managed to ‘harvest’ the images we needed.” In statements accompanying the film’s release, Xu compared his studio’s working method to that of the global, cell-phone-activated taxi service Uber, saying that one doesn’t need to own the cars or cameras, one simply needs to control the network. 

There is a consistency here with Xu’s artistic practice of the last three decades, which has frequently appropriated creative modes outside of contemporary art and explored themes of language, fakery and meaning. His previous works include a 604-page, four-volume book series, composed of 4,000 fake or meaningless Chinese characters, A Book from the Sky (1987–91); a system for writing Chinese calligraphy using the Western alphabet; and a 112-page novella composed of emoticons, Book from the Ground (2003–14). Like Dragonfly Eyes, Book from the Ground exists in a non-art commercial context; it was published as a book series with an ISBN number, and is for sale on Amazon.com. As a story, it is weak, because Xu’s real interest is in the reinvention of language, not creating compelling narratives. As he explained, “To discuss today’s issues, I need to use today’s ways of talking. I can’t find that in any preexisting system.” In venturing further into the quasi-human or transhuman language of networked surveillance in Dragonfly Eyes, he suggested, “We came to a new understanding of the so-called boundaries of reality.”

The surveillance state, one should note, has special implications in China. Beijing is currently implementing facial recognition software on a massive scale, using chatbots to delete politically sensitive posts off social media, and will roll out a “social credit score” for all citizens by 2020. This last system will assess financial history and online behavior, including social-media posts, and potentially limit real-world freedoms. The macro and micro implications of these developments are wide-reaching. If the horror of George Orwell’s omnipresent government in his dystopian 1984 has in some ways finally come true—we are now surrounded by a ubiquity of cameras and screens, with which we are in constant interaction—technology critics also warn that big data will undermine individual freedom in democratic states. 

Xu, the former vice president of Beijing’s Central Academy of Fine Arts, is likely to keep any criticisms of his own government deeply veiled. But one could also argue that if Dragonfly Eyes is in fact a political critique, it is not directed against any one political system, but rather against technology itself. 

In that sense, the film is even more disturbing than Hollywood’s latest vision of technological dystopia, Blade Runner 2049, which, although gripping and majestic, ultimately fails at any real critique of technology. In Blade Runner 2049, the new technologically enhanced beings that populate its story—replicants and AIs—all act exactly like humans and not in the least like machines. The Hollywood film fails to imagine anything like a non-human, technological intelligence, which is precisely what Xu Bing and his team, sitting in a small artist studio and tapped into a global surveillance network, have revealed—an otherness in the way computers see and process the information of physical human reality.

SUBSCRIBE NOW to receive ArtAsiaPacific’s print editions, including the current issue with this article, for only USD 85 a year or USD 160 for two years.  

ORDER the print edition of the March/April 2018 issue, in which this article is printed, for USD 15.