4.1 The primacy of language
A defining feature of a sophisticated society is how it communicates with humans, machines, and institutions. That is how humans control and coordinate strategy. But the relationship between language and power is intricate. Thoughts get communicated through language (speech, script, and sign) and emotion (body language). Benjamin Lee Whorf (1897– 1941) said, “Language shapes the way we think, and determines what we can think about.” And Ludwig Wittgenstein (1889 –1951) said, “The limits of my language mean the limits of my world.” Thus, whoever speaks depends on language, but ultimately the power of language lies not with the speaker but with language itself.38 Anyone can acquire the power of language, even AI machines.
Communication channels for the millennials expanded suddenly with the invention of the microchip in 1959 (a product of the electronics revolution) that enabled the personal computer (1975 as kits) and eventually the present day ubiquitous smartphone that include high-speed mobile broadband 4G LTE, motion sensors, camera, and mobile payment features that fit into a shirt pocket. Consequently, the world of the millennials is much less human-centric than machine-centric. STEM educated millennials need, not just a language with which to express myths but also logos, i.e., a language with which to reason in accountable writing and speech. This is fulfilled, to a good degree, in esoteric (and hence limited to relevant experts) modern scientific languages, by heavily depending on the language of mathematics. Scientific languages speak with “the highest, universally binding authority, world-wide about everything in the world” and its “authority is fundamentally egalitarian and democratic; for it and with respect to it, nothing counts but ‘the non-violent force of the better argument’ (Jürgen Habermas).”39 This, of course, does not eliminate the criticism that scientific languages also narrow man’s perception of reality to what can be expressed in scientific language. Our successor species may evolve to deal with this problem better because their survival may depend on it.
The electronics revolution was ignited by developments in human-computer interface, networking of digital machines, and microchips. In the millennials’ world, man-machine interaction and action-at-a-distance with the speed of light is taken for granted. That is only the beginning. Globally connected digital devices, the Internet of Things40 (IoT), with each device becoming smarter by the day with embedded AI that include cognitive functions, is on its way to becoming ubiquitous. Along with it, human-computer interactions are becoming more like human-human interactions because AI has now made significant advances in mimicking cognitive functions. This raises the specter of en masse retirement of humans even from “intelligent” tasks. Technology and society are now so intricately interwoven that without understanding their relationship in terms of information flows, humans may inadvertently place themselves in extreme danger of becoming unemployable. For developing countries like India, the danger is even more.41 Emerging technologies like IBM Watson42, Google’s AlphaGo43, and Carnegie Mellon University’s Libratus44 clearly indicate such a possibility may occur soon. On another important front, AI may provide an amazing service, that of peer review of scientific papers, in the near future.45 When such a feat is accomplished, it will be but another step to robots writing papers making even human researchers redundant, inefficient, and obsolete. As Janne Hukkinen noted, “New knowledge which humans no longer experience as something they themselves have produced would shake the foundations of human culture.”46 Greater the sophistication in communication, the more elaborate will be its language, especially its associated grammar. The post-industrial economy will depend on machine interpretable, error-free communications.
When talking about AI and its impact on jobs for humans, we tend to overlook that AI is human-created, and the related technology is often patent protected. It is this technology that is super-charging AI robots into delivering intellectual output. We now face an unusual situation. By law, patents can be granted to only human inventors of a novel, useful and non-obvious invention and they must describe their invention in writing (written description) with clarity and in sufficient detail so that others knowledgeable in the arts related to the invention can reproduce it independently. The law does not permit patents to be granted to machines that invent. Further, if an AI machine comes up with an invention, it will be considered an obvious invention, because similar machines or its clones will be deemed able to come up with the same inventions if called upon to do so! It will be obvious invention by machine instinct. Furthermore, given a powerful enough AI machine, it will often be possible to show that a patented invention could have been invented (and hence the invention anticipated based on prior art) by this machine if only it had been asked to create one with its current and possibly past knowledge and computing power.
Historically, the intellectual property (IP) protection system arose to facilitate, indeed drive, economic growth as a means of improving people’s well-being by equitably rewarding inventors. If someday AI-driven humanoids rule the Earth, economic growth and human well-being may no longer be a criterion for generating IP. It will be generated on its own, algorithmically or randomly, all, of course, under themathematical laws of Nature. Thus, such studies as conducted by the National Academies (of the U.S.) may well be futile efforts in ‘Advancing Concepts and Models for Measuring Innovation’47
The IT revolution has unevenly affected the world in terms of enhancing living standards, governance, the economy, employability, etc. Successful diffusion and adoption of IT is time and resource intensive. It has led to increasingly uneven productivity gaps between frontier firms and others. Income and wealth inequality has increased. In the past two decades, the top one percent have gained enormously, while the bottom 80 percent have steadily lost. The job-mix in the economy continues to change faster than people can adapt since many routine tasks are being automated; skill levels required for new jobs is steadily rising. There is great uncertainty about future wage-earning jobs– what will they be, how long will they last, what skills will they require, what future prospects will they carry, where will they be available, how can one reskill, etc. In fact, how the world will reorganize itself under the pressure of technological advances is not at all clear. There will be phase transitions galore at different skill levels for humans. Gathering and analyzing data to even understand or discern trends as to what is happening around us has become impossible because the situation is changing so fast. Statistical analysts are out of their depth. New ways to integrate various data sources without compromising privacy and confidential business information could reveal useful information about the changing workforce but it would only be useful as history and not for planning. The best forecaster for the future remains Kurzweil.48 The situation is heading towards chaos. However, education will be a key influence on worker income, but matching education with opportunity will be a loaded game of dice.
4.2 Law of phase transition
Languages (natural and programming) connect humans and machines in seemingly random fashion. Graph theory tells us that massively connected men and machines will lead to phase transitions. Everytime connectivity breaches a critical point, a phase transition occurs in the way nodes form or reform into clusters. The millennials’ world is already vastly different from the world of their parents who were far less connected. Since 1960, we know from graph theory that if we have a set of $n$ nodes and start linking them randomly, then when $m = n/2$ links are made, a phase transition in the graph occurs in the sense that a giant connected component in the graph spontaneously appears, while the next largest component is quite small.49 Additionally, such giant components remain stable in the sense that how we add or delete $o(n)$ edges, the size of the giant component does not change by more than $o(n)$. What is seen is that even uncoordinated linking, whether protein interaction networks, telephone call graphs, scientific collaboration graphs, and many others show such generic behavior of forming giant components.50 This immediately suggests a basic involuntary mechanism by which a society at various levels of evolution spontaneously reorganizes itself as nodes (people, machines, resources, etc.) connect or disconnect in apparent randomness. The effect is highly visible in the IoT world of the millennials who become spontaneously polarized on issue-based social networks.
Rapidly increasing connectivity among men and machines has already imposed upon the global socio-politico-economic structure, a series of issue-dependent phase transitions. More will occur in areas where massive connectivity is in the offing. Immediately before a transition, existing man-made laws begin to crack, and in the transition, they break down. Post-transition, new laws must be framed and enforced to establish order. Since such a phase transition is a statistical phenomenon, the only viable way of managing it is to manage groups by abbreviating individual rights and enhancing those of its leaders. The emergence of strongman style leadership and its contagious spreading across the world is thus to be expected because job-starved millennials will expect them to destroy the past and create a new future over the rubble. It appears inevitable that many humans will perish during the transition for lack of jobs or their inability to adapt to new circumstances. Robots will gain dominance over main job clusters while society reorganizes. Ironically, robots neither need jobs, nor job satisfaction, nor a livelihood. So there will be an aura of rational ruthlessness in the reorganization.
4.3 The fragile limits of human intelligence
We posit a definition of knowledge and intelligence. Given massive amounts of data, condensing it into a smaller sized axiomatic system from which a subset of the data can be regenerated, and data outside the subset can be created by interpolation and extrapolation. The resulting axiomatic system is called knowledge, and the process of deducing the axiomatic system is called intelligence. The process of regeneration, interpolation, and extrapolation is called computation. The baby steps of knowledge creation are what babies take with a primitive but malleably evolvable brain–the physical matter that acquires, analyses, and processes data. Further evolution of the brain depends on the environment with which it interacts by establishing multiple feedback loops by continuously condensing its already acquired knowledge and knowledge acquired from the environment into abstract concepts, which themselves undergo further compaction.
[ 38 ] Weiß & Schwietring (2018).
[ 39 ] Weiß & Schwietring (2018).
[ 40 ] IoT is the network of physical objects–devices, vehicles, buildings, etc., which are embedded with electronics, software, sensors, and network connectivity, which enable these objects to collect and exchange data.
[ 41 ] Yahoo (2016).
[ 42 ] Best (2013); Young (2016).
[ 43 ] Gibney (2016).
[ 44 ] Solon (2017).
[ 45 ] See, e.g., GenomeWeb (20170221). See also: Stockton (2017).
[ 46 ] As quoted in Stockton (2017).
[ 47 ] NAP (23640) (2017). See also: Crawford & Calo (2016); White House (2016a-c).
[ 48 ] Kurzweil (1999), Kurzweil (2005).
[ 49 ] Erdős & Rényi (1960). See also: Krivelevich & Sudakov (2012).
[ 50 ] Bollobas, Janson & Riordan (2007).