“Upon my soul we might have known that some day or other the Robots would be stronger than human beings, and that this was bound to happen, and we were doing all we could to bring it about as soon as possible.”(Busman, R.U.R, III-47)
In literature and cinema, artificial intelligence and singularity intertwined more often than they don’t wether it is according a benevolent or disaster narrative. When in London I finally got the opportunity to see the highly anticipated exhibit (at least by me) AI : More Than Human at the Barbican Center. Described as a “ ‘festival-style’ exhibition [hey you have to attract the masses!] […] Bringing together artists, scientists and researchers, this interactive exhibition offers an unprecedented survey of AI with which you are invited to engage head-on”.
I was already intrigued by the name itself AI : More Than Human. Why this more between AI and human ? Was AI superior to human ? Were men to become more than humans in a near future ? Questions which for now remained unanswered. As I didn’t want to spoil myself by tracking the hashtags of the show on social media nor read reviews, I only stuck to those lines before going. Plus it was an opportunity to appreciate the Barbican once again. The center itself is an experience : The gloomy buildings cut in concrete are the ideal symbol of Brutalism. Getting lost in this forest of infinite towers and bars, be seduced by the unlikely post-apocalyptic courtyard and its turquoise water-flows are part of the show. As I got swallowed by these masses of stones, the sun disappeared and I was soon welcomed by the infamous fresh air and the darkness of the Barbican vast corridors on that early summer afternoon. The immersion could start. This is actually the perfect word to describe the show.
Indeed, immersion in arts has become a recurrent theme. Visitors want to be included and take part in wether it is by having their picture taken in front of an art piece – and no longer take a picture of the piece itself – , Instagram-Material pieces or immersive experiences.
Thus the Barbican created a whole experience based on immersion and engagement almost in a gaming/education-style. Lots of definitions, timelines, and fun facts. And tons of interactive platforms and stations waiting for nothing else but visitors to play with ! So what did we see there ?
The blow of the exhibit was probably ALTER3, a Japanese robot developed by Hiroshi ISHIGURO and Kohei OGAWA. ALTER3’s hands moving frenetically spark some curiosity. The disturbing apparence of the robots : a bare machine body with a human-looking face wasn’t the only thing to be amazed by. Using deep learning, the robot learns through the interactions with its surrounding and experience.
We also met Mitsuku – which despite its name is not Japanese – the best performing chatbot developed to date created by Steve WORSWICK. There was a great focus on progress made by facial recognition but also impact on speech and language can AI have on society. Washington University played with several speeches of OBAMA, synthesised them to create a whole new ones with opposite meaning.
The effects of automation in language and the thought process were also explored and neutral thinking hinted. NY Times has been testing and training a system to moderate the comments section of its websites for example. Ultimately, the system will be able to not rely on human assistance to sort out comments and accept or ban some.
Singularity is a largely discussed and polarising topic when we evoke AI. Singularity is the narrative according which a day will come where artificial intelligence will awake after years of perfectionism and develop its own mind marking the end of humanity and the beginning of a post-human era.
The term was coined in 1958 by mathematician Jon VON NEUMANN recollects his friend scientist Stan ULAM. By that time Von Neumann was referring to normal progress not involving superhuman intelligence however.
The imminent creation by technology of entities with greater than human intelligence, the awakening of artificial intelligence machines and computers with an intelligence way superior to humans were then developed the following decades as computer development progressed.
1993 is a landmark as mathematician Vernor VINGE (San Diego University) came up with The Coming Technological Singularity: How to Survive in the Post-Human Era in which he states “Around year 2023, we will create a superhuman intelligence that will sets off the end of the human era”. Point. Blank. Period.
According to Vinge singularity is only a matter of time since it mostly depends on progress made in hardware performance already exponentially improving. Computers and large computer networks will then awake to become those “superhumanly intelligent” entities. Another forecast is the widespread of computer/human interfaces enhancing humans natural intellect abilities. If Vinge got the date of the accomplishment of his prophecy wrong, some outlook and forecasts on AI are accurate and time has not proved him wrong yet.
For example he predicted a wave technological unemployment with more and more machines replacing higher level jobs.
“An ultra intelligent machine will be built and that will be the last invention than man need make”
Vinge comes to the conclusion that greater-than-human intelligence will drive progress, that progress will exponentially accelerate and produce still-more intelligent entities on a still-shorter time scale.
Man is able to create simulations of the world in the mind to anticipate and solve issues ahead of the natural selection process, differing from animals. By analogy artificial intelligence is able to create simulations way faster than humans can do. Singularity as seen by Vinge is the “point where our old models must be discarded and a new reality rules”.
Far more ahead of his time was British mathematician I.J. GOOD, the man behind the concept of “Intelligence Explosion”. In 1966 he prophesied that a day will come “Within the 20th century, an ultra intelligent machine will be built and that will be the last invention than man need make” According to the mathematician, we may one day design a machine that surpasses human skills at designing AIs. After that, this machine will improve its own intelligence faster and better than humans could, which would make it even more skilled at improving its own intelligence.
“Since the design of machines is one of these intellectual activities, an ultra intelligent machine could design even better machines” he said. Intelligent explosion is the pragmatic take on singularity.
Machines not needing man anymore and awakening has been a recurring pattern throughout the 19th century until today in literature and cinema starting with Shelley’s Frankenstein (1818). Since then, we’ve seen lots of computers plotting behind our backs while we were asleep, man-made creations gone mad against their equally mad scientist creator, and still more human-looking humanoids trying to kill our favourite hero to replace them.
The prospects of AI reaching singularity have long be explored since the early 19th century matching the spread of the Industrial Revolution culture (industrialisation of mankind ?)
One work we cannot not mention is Czech play R.U.R by Karel ČAPEK first published in 1920. R.U.R – Rossum’s Universal Robots – is important as it marks the first occurence of the word robot to refer to humanoids and androids. Robot derives from Robota which means “forced labour” when Robota derives from rab which means slave. Karel Čapek credits his brother Josef for the finding.
R.U.R is a very fascinating play to read considering how ahead of its time it was.
Domin, runs a factory producing human-looking robots exclusively designed for chores and labour he inherited from Old Rossum who first discovered a special cell to design robots. Years goes on and R.U.R becomes world-wide successful company which has subverted nations. And a robot-based economy has arisen. Nations are ordering more and more robots especially to strengthen their defence and have more soldiers. Also a decline in natality is seen across the world as we need less human workers so people. This is when robots awaken and threat humanity. The play ends on a quirky plot twist I’ll let you discover.
Despite being mentioned on rare occasions, R.U.R can be seen as a cornerstone for many science fiction and literature works and movies such as Fritz LANG’s Metropolis (1927) which also explores similar paths : future, man versus machine, master versus worker. R.U.R is also about social engineering. The subtext of the play is the dehumanisation of men into literal robots during the Industrial era. See this quote of DOMIN (I – 29) talking about robots :
DOMIN : What sort of worker do you think is the best from a practical point of view?
HELENA : Perhaps the one who is most honest and hardworking.
DOMIN : No; the one that is the cheapest.
Here robots can easily be replaced by workers. Similar to this quote from Metropolis when a worker talks to Metropolis’s tycoon heir Freder.
“The machine ! Someone has to stay at the machine !”
Those early 20th century works are strong critics of Fordism. Fordism which is ruled by principles the most known being the assembly line, the standardisation and uniformisation of products (over craftsmanship) and the scandalous rise of wages to allow people buying the products they may have contributed producing. Thus, the awakening of machines can be seen as a metaphor of workers revolutions emancipating themselves from slavery they’ve been put into and finding back their mind. Those sci-fi and dystopian productions tell more about the human nature (or the loss of it) than the actual alleged desire for domination of machines on mankind.
“ Humans have long dreamed of superhumanity and AI appears as the ultimate servant… as long as it serves humans purpose. ”
Back in our times, the same disruption caused by the Industrial Revolution and the Fordism could happen again with AI. It will. It is, actually. In Homo Deus (2015), Yuval HARARI talks about the “Gods and the Useless” just like unskilled workers living underground and the elite living above in Lang’s Metropolis. In this present-future, we’ll need more managers than actual workers creating a thicker gap between the ultra qualified managers and engineers and unskilled workers, between the winners and losers of this AI-based economy. French surgeon Laurent ALEXANDRE reports that “AI transforms social organisation and favours intellectual elites versus the people […] devalues handcraft and the middle-class”.
Humans have long dreamed of superhumanity and AI appears as the ultimate servant… as long as it serves humans purpose.
Singularity, Intelligence Explosion, are we even close ?
“It’s still one of the greatest mysteries how this three-pound piece of matter can give us all our subjective experiences, thoughts, and emotions”Dr. Heather Berlin
Depends on how you define intelligence. Human beings have the ability to solve problems in a given environment wether they have already experienced it or not. It is part of the general intelligence along with emotions, thoughts, creativity. This is actuality not the case for machines. Machines they surpass humans abilities in numerous fields they’ve previously been trained for. A machine cannot solve a problem if it has not been trained for before. This is the first – and crucial – difference between humans and machines. Also, machines are unable to understand the principle of emotions and develop some sort of thought. It is an unintelligible principle to the machine for now.
A chatbot like Mitsuku can chat with a user and reply according to the tone of the conversation but it doesn’t mean it does understand the meaning of it. A chatbot will use answer A and not B because it has previously learned it was the appropriate option to question A. As a matter of fact Worswick is currently adding abilities to Mitsuku so it can fight back abusive content it receive (accounting for 30% of the content received.) Worswick wrote a column on Medium discussing wether chatbots should develop those skills or shall he or others chatbots makers simply rule out offensive content which may be downgrading the performance of their creation.
In that way, machine cannot surpass the human intelligence despite already surpassing us when it comes to the speed of calculation, data search and organisation, probability and algorithm generation. Those abilities solely depend on the hardware performance. Vinge predicted we will one day design hardwares close to perfect in term of speed but that will be the end of it. Luke MUEHLAUSER (2013) comes to the same conclusions. He suggests we will perfect AI but not to the point of developing something close to the human mind. DREYFUS (1972) and PENROSE (1994) argue that human cognitive abilities will never be in the abilities of a machine.
“Commercial digital signal processing might be awesome, giving an analog appearance even to digital operations, but nothing would ever “wake up” and there would never be the intellectual runaway which is the essence of the Singularity.”Gustav STENT , The Coming of the Golden Age (1969)
Futurist Ray KURZWEIL (2005) predicts that machines will reach our level of intelligence by 2030 creating “a profound and disruptive transformation in human capability” by 2045 supporting Good’s intelligence explosion instead.
Finally and I’ll leave it there, 4D explorers strongly believe we on a daily and individual basis are strengthening AI which only desires one thing : collect enough knowledge and data to reach the Singularity. For now this AI would only be a vast virtual nebula that when awake would materialize by connecting all devices and smart networks together in order to rule us. Artificial Intelligence replicates nature’s pattern where 1+1=1 and all information collected by one device is fully saved and processed by other device creating one single “mind” . In this scenario the ascent of 5G would fasten this as data flows and information still more rapidly and extend AI exponentially. A theory to place on the side of fantasies as it is not backed by evidences.
From AI to IA
Transhumanism appears as the latest agenda pushed when it comes to AI. Indeed if for now AI developing its singular mind seems farfetched (let’s be honest), it could be the tool to enhance human natural intellect and. By taking the best of both we will finally develop superhumans as many dream of, moving towards transhumanism.
Vinge talked about IA (Intelligence Amplification) which is what most scientists and mathematicians tend to align nowadays. Machines could be used as assistant to the human to perform well defined actions. For now biological cognitive enhancement and brain-computer interfaces (meaning directly connecting to the brain) have both been tested and were successful in improving general intelligence. Biological science may provide means to improve natural human intellect.
Moving towards transhumanism means no more man versus machine but man and the machine.
Evidences of AI reaching singularity in a near future are as heavy as evidences prophesying the opposite. Indeed much of the singularity seem confronted to mind. Only limit is of the mind and emotions which are quite unique.
On the other hand argue that even if a day might come when AI will perfect its hardware, learning and algorithmic abilities , it still won’t be able to compete with the humain brain and be able to develop some sort of mind and emotional intelligence. Indeed can emotions be trained ?
As for now, AI remains for the general publique playful tools and humanoids like Mitsuku the chatbot, Sophia the first humanoid-citizen and ALTER3. Interesting to see how the narrative about AI powered machines and devices has evolved in order to gain back the trust of people and put on a more friendly image. AI is here to assist humans, enhance our abilities and is no threat is the new adage. Amazon’s Alexa and Echo or Google Home are not called voice assistants for nothing. Also AI is being introduced to the mainstream via culture. Sophia the humanoid has been granted citizenship and is covering magazines, The MET gala explored the theme in 2016 with Manus x Machina. More and more museums are integrating digital arts and AI-focused installations into their agenda. Last year Paris Grand Palais welcomed Artists and Robots during spring while Fondation EDF is exclusively dedicated to arts and social experiments using data and new technologies such as VR and AI. The movie H.E.R explored the notions of love and AI while Black Mirror regularly dig deeper into the drifts technology bring us. Finally, I saw last week a billboard in the underground announcing the arrival of the Kara and Connor, the first androids wax doll at Musée Grévin (French wax dolls museum). Kara and Connor are actually the main characters of video game Detroit Become Human which has an interesting synopsis.
Singularity is a fascinating theme as it universally challenges the notion of progress which has been the West religion since the late 19th century. How much progress is too much ? When do we reach posthumanism ?
Across the past two centuries, the narrative about AI has evolved and drastically changed. AI first occurred as creatures and abominations (Frankenstein) during the 19th. It then appeared as dumb machines made to serve humans. However mid-century the narrative evolved again and we explored superhuman robots first admired then feared. Where are we now ? As I write those lines, we can clearly see a drastic change of narrative which began in the 2010’s and should be named Life with AI Man and AI shall coexist and merge in total equality. This is how we will reach the ultimate singularity. Is modernity worth the loss of humanity ?
Laurent, Alexandre, “L’Intelligence Artificielle produit des gilets jaunes”, in L’Express, November 2018
Karel, Čapek, “R.U.R”, 1920
Good, Irving John. 1965. “Speculations Concerning the First Ultraintelligent Machine.” In Advances in Computers, edited by Franz L. Alt and Morris Rubinoff, 31–88. Vol. 6. New York: Academic Press, 1966
Luke, Muehlauser, “Intelligence Explosion FAQ” , Machine Intelligence Research Institute, 2013
Vernor, Vinge, “The Coming Technological Singularity: How to Survive in the Post-Human Era.”, In Vision-21: Interdisciplinary Science and Engineering in the Era of Cyberspace, 11–22. NASA Conference Publication 10129. NASA Lewis Research Center, 1993