Ushering in the Era of “Beneficial Intelligence”

The notion that there will be a time when technological change will rapidly accelerate, transform, and then merge with human life — in other words, the Singularity — has been with us since at least the fifties. Concepts of non-human entities fashioned in a kind of human image, such as the Golem from Jewish literature, have been with us even longer. By the time the term “robot” entered common parlance, we had already collectively imagined a future full of machines taking on human characteristics and forms. How close are we to the realizations of such concepts, and how much influence do we have in the shaping of the future?

Rabbi Loew and Golem by Mikoláš Aleš, 1899.

A contemporary smartphone is computationally more powerful than the first Space Shuttle, and today it is unremarkable to own and use one to communicate, calculate our position in space, transmit sounds and images. Over the past hundred years, we have moved from just naming what a robot is or could be, to having ubiquitous, powerful computing technology at our fingertips. The rapid acceleration that Ray Kurzweil spoke of in his popular book The Singularity is Near is palpably a part of our present, and the notion that the Singularity is on the horizon no longer seems an improbable idea, limited to science fiction.

How we’ve chosen to approach the Singularity through the medium of popular culture often reflects a paranoid, dystopian view on the Singularity, tapping into fears of losing control and the primacy over the Earth that we have established and now call the Anthropocene. Though Kurzweil has a more nuanced approach and claims that the Singularity is “neither utopian nor dystopian”, it proves irresistible for us to imagine an almost fetishized vision of submitting to our more intelligent cyborg cousins. Humor has begun to trickle into this otherwise somber discourse — it was recently announced that low-brow nerd comedy star Seth Rogen (of Knocked Up fame) will direct a sitcom about the Singularity, for example.

One of our big pre-Singularity human failings, however, is the incredible difficulty we face in imagining the future without the baggage of the present. Our vision is limited, and is often disappointingly shallow and cosmetic. When for example the virtual world of Second Life opened up in 2003, the possibilities on how to represent ourselves and our world were vast. In the face of such choice, most users gave themselves a makeover in the style of our current Barbie and Ken beauty standards. American Apparel and other mainstream retailers set up shops. Given the options, we recreated a shopping mall. Not long after, vast tracts of virtual land were abandoned and in typical human style, Second Life filled up with garbage. Are we truly capable of envisioning the Singularity and post-Singularity if we populate a virtual world with avatars and objects that make The Jetsons look like a radical vision?

Abandoned Ohio University campus in Second Life. Screenshot by Patrick Hogan.

Fortunately, the whole idea of the Singularity saves us from having to imagine much beyond our tiny, blinkered vision: as Stephen Hawking has commented, Artificial Intelligence will eventually result in “…machines whose intelligence exceeds ours by more than ours exceeds that of snails.” AI will be so far ahead of humanity, we won’t need to strain for interesting ideas, because AI will have them for us.

Waiting until such time that we can fully defer to our soon-to-emerge technological superiors is a cop-out, however. Not to mention, we still have yet to grasp how our hidden biases shape the tools we create. We are slow to acknowledge that, for example, the algorithms we write perfectly demonstrate our own flawed thinking. Racial and gender bias has been proven in our seemingly objective algorithmic creations. Without taking care, our post-Singularity lives will be just as fraught with irrational and damaging bias, because we baked it in. That the AI which will replace us is smarter than us is not a reassurance. Some of the most reviled leaders of human history acted unethically, but could hardly be called stupid, so “smart” is no salve to problems like inequality or discrimination. The ethics of AI behavior in a post-Singularity world is a legitimate concern.

The concept of a post-Singularity is meant to push our buttons: or induce an eye roll even, at the use of the prefix “post” once again. Deliberately placing it there to provoke us, we are encouraged to think about not just the event of the Singularity itself and how we get there, but what can we control before we lose control? It’s imperative to find ways to address the political and social consequences before they happen, and not allow our thinking to stop at the event horizon itself. It would be a huge failure to assume that we can build the capacities for AI and simply let it run because it will be smarter than us. Our creations inherit our flaws and hang-ups. Or as Stephen Hawking also remarked, it’s critically important to “…shift the goal of AI from creating pure undirected artificial intelligence to creating beneficial intelligence.”

This “beneficial intelligence” that Hawking refers to will not be easy to implement (as his own status as a casually sexist physics genius reveals). Reflecting upon the sluggish development of a philosophy of technology, Langdon Winner remarks: “Why has a culture so firmly based upon countless sophisticated instruments, techniques, and systems remained so steadfast in its reluctance to examine its own foundations? Much of the answer can be found in the astonishing hold the idea of “progress” has exercised on social thought during the industrial age.”

Relentless technological progress inevitably leading to the Singularity heads in one direction, and ethical considerations, legal reforms, and safety planning cannot keep up. As noted with the Second Life example, our creativity and inventiveness often fail us when presented with a blank slate or a long-term challenge. For example, dealing with our own waste in the long-long-term reveals how difficult it is for us to think on such a scale. One major challenge facing humanity is our management of nuclear waste. Several ideas have been posed for how to warn future human beings about the presence of securely buried nuclear waste. Waste containment sites must remain undisturbed for 100,000 years, and given that daunting timeline, proposals for architecturally forbidding landscapes and wordless diagrams discouraging digging have been proposed to protect these sites from human interference. An idea arose in the planning of a Finnish containment facility: perhaps the waste should simply be buried very deep with no external marker to attract attention? This proposal might cast a cynical regard on human nature and our ability to restrain our curiosity, but perhaps this strategy is the best bet. Would our post-Singularity AI leaders be able to show more restraint, or would they be intelligent but arrogant, with the failings of their human ancestors embedded too deep in their programming?

Trailer for Into Eternity

Artists are ideal agents to more deeply excavate of this idea of the Singularity and beyond. Free from a uniform agenda, artists and designers can imagine where we can go and question our vision at each step. For example, the work of artist Mike Pelletier presents beautiful though disturbing distortions of the digitally-captured human body ultimately cause the viewer to ask: why do I find this unsettling? His work posits a type of possible mutation of anthropocentric beauty into something more dynamic.

Time of Flight by Mike Pelletier

Artistic pioneer Remko Scha, who believed that art could be created by machines, demands that we conceive of our algorithms and machines as equal partners, as collaborators. Scha’s work suggests to us that what we might see post-Singularity is more strange and oddly beautiful than we can currently conceptualize. This kind of thinking brings us closer to an appealing approach to the idea of post-Singularity: to give up control but instead of imagining the worst, anticipating pleasant surprises and finding the post-Singularity as a potential field for aesthetic revolution and new forms of beauty.

Over time, our influence will dim as AI creates art that another AI entity would appreciate, and the human witness or viewer becomes less important. The reduction of human influence until it is nearly invisible is articulated clearly in the piece Shiv Integer by Matthew Plummer-Fernandez and Julien Deswaef, wherein a bot finds and combines online 3D models into unique new forms. The resultant shapes are visual slapstick, and an almost mocking outlook on what algorithms might make when left to run on their own. Soon enough though, those techniques will become more sophisticated.

The way in which an algorithm uses found 3D models has its mirror in a case recounted by Benjamin Bratton in the 2015 Benno Premsela Lecture. Bratton began his lecture by describing a futuristic pod city complex built in Taiwan in 1978 which only functioned for two years before being abandoned. When the time came to demolish it, it was discovered that there were thousands of an unknown species of insect, the orchid mantis, inhabiting the ruins, for reasons unknown. A species of orchids (for which the mantis was named) also flourished there in the ruins left behind by humans. This historical case underscores the notion that we are impressive builders, but we are also in competition with dozens of other known and unknown species on this earth — some of whom will be our own creations, post-Singularity.

2015 Benno Premsela Lecture by Benjamin Bratton

Bratton continued his lecture by saying that he believes the Anthropocene will be short-lived, in his words, a mere “geopolitical instant”. Bratton states: “Human beings are vanishing. Our cities are not our own. We are building habitats for other forms of life.” Referring to the mantises in the Taiwanese pod city, he continues: “We are the robots for future insects.” Bratton suggests that this is not a failed future, but a successful future — wherein we provide infrastructural support for future other species.

Suggesting that our moment of dominance may be fleeting, even though this is presented as a kind of success, Bratton taps deep into our fundamental fears of losing relevance and being completely wiped out. As we see in one of the many variations on the legend of the Golem, we intend for our creations to protect us, and when they begin to lose control and become destructive, we know (as we are their creator) how to stop them, so there is a safeguard against further harm or loss of human control. As we begin to build things we may only partially understand, or things which may have purposes we could not foresee, we lose the ability we wished for with the Golem: the control lever to stop the situation from escalating.

One of the first known uses of the term robot was a play by Czech playwright Karel Capek called Rossum’s Universal Robots. The play depicted a scenario which has now become a cliché in popular culture: the robots revolted against their human overlords. The word robot came from the Czech word robota, which means “serf or laborer”. The post-Singularity conflict we are imminently facing is more complex than the rising up of a machine class built to be slaves, or a being, the Golem, summoned from clay with an inherent, exploitable flaw. As Langdon Winner put it, “…technologies are not merely aids to human activity, but also powerful forces acting to reshape that activity and its meaning.” Technology and culture make meaning together, and in what could be our waning years as human rulers of this planet, we have a moment to take stock. It’s time for a reckoning on how effectively we have marshaled technology and culture as forces for positive change and human evolution, before we face an unprecedented kind of robot rebellion, wherein our machine servants — who have learned all our bad habits — are finally able to outsmart us.

Text commissioned for the 2016 edition of the GOGBOT Festival: Post-Singularity, Enschede, The Netherlands.

Sources:

  • Bratton, Benjamin. Benno Premsela Lecture, Het Nieuwe Instituut, Amsterdam. 1 November 2015.
  • Golem of Prague. Project Gutenberg
  • Hawking, Stephen. Ask Me Anything (AMA) on Reddit
  • A History of Robots, from Science Fiction to Surgical Robots. Hockstein, N.G., Gourin, C.G., Faust, R.A. et al. Journal of Robotic Surgery (2007)
  • Into Eternity. Film, 2009. Trailer
  • Kurzweil, Ray. The Singularity is Near. Viking, 2006
  • Langione, Matt. Will Art Save our Descendents from Radioactive Waste? Jstor.
  • Winner, Langdon. The Whale and the Reactor. University of Chicago Press, 1986.