- I am very happy to be here at the opening of this prestigious event, which gathers together the best young science minds from around our region, and exposes them to discussion and debate, and to learning from some of Asia’s finest senior scientists. These include highly esteemed Nobel Laureates, among whom I feel very honored to be speaking today.
- This is an extremely exciting time to be a young scientist, as you are, in the many areas covered under the rubric of STEM (Science, Technology, Engineering and Mathematics), the theme of this year’s science camp. It may possibly even be the most exciting time ever to be a scientist, as ever-more astounding advances are being made in so many fields. Science fiction is becoming science fact,[1] as flying cars and bionic limbs become a reality. And advances in artificial intelligence and robotisation are pushing boundaries in directions that are both expected and unpredictable. The speed and extent of change is such that some futurists believe that we are reaching an historical tipping point, which will transform our world beyond recognition.[2]
- This creates extraordinary opportunities for the world, and for yourselves, as both the beneficiaries and architects of some of these advances. Some of you here attending this camp may well go on to shape some of these revolutionary technologies. But the acceleration of technological change also presents major challenges for us all, as we navigate through this turbulent and unpredicatable world. Advances in AI (Artificial Intelligence), robotisation and bio-technology are already allowing innovations such as genetically modified insects and artificially enhanced or manipulated humans. These are very welcome for their medical and humanitarian contributions. But such developments also raise deep and challenging questions about even what it means to be human.
- Beyond such difficult philosophical questions lie the more immediate and equally challenging issues of the future of work and the workplace in an era of increasingly capable AI and robotics. The impacts of recent advances will probably be felt most quickly in this area, driven by the economic logic of investment for profit. This area also has very direct implications for yourselves, whose working lives may look very different from those of previous generations. This transformation of employment patterns will also have broader political implications, as society tries to adapt to the new realities. The economy will continue to evolve regardless, along with technological innovation.
- There are many dire predictions of the likely impact of the coming wave of technological change on employment and the workplace.[3] One study suggests that at least 50% of seven hundred different current job categories may be fully automated in the coming years.[4] At the same time, large productivity gains will be generated by the new technologies, in the same process of value-added through specialization described centuries ago by economist Adam Smith. But the costs and benefits of the coming transformation are unlikely to be evenly distributed, and this could contribute to worsening inequality. Combined with mass unemployment, this could result in politicial and social upheaval on a massive scale, as depicted in dystopian science fiction movies.
- It is also here, however, in the social and political realm that we can work to influence these technological and scientific advances, and regulate them effectively. In this way we can help to ensure that their power is exploited for the greater good. Scientists such as you can and should play an important role in this area of policy development and the regulation of scientific and technological advances. By applying your specialist knowledge, you can help to ensure that these advances are effectively controlled and managed for the benefit of all.
- A useful contribution can also be made by transnational scientific exchanges and debates, including gatherings such as this one. The challenges presented by these coming changes must be examined more closely and discussed more openly. The more we understand what is happening, the better we can control it and channel it effectively. High quality and more widespread education in the STEM fields is a central aspect of this process. Critical thinking and problem-solving skills, along with creativity, will only become more important.
Ladies and Gentlemen:
- It is in this area of the economic and political impacts of technological change that I want to focus my remarks today, looking particularly at the world of work. This area highlights some of the more radical ways in which these changes could play out.
- The transformation of the workplace will affect the whole range of jobs from ‘blue’ to ‘white collar’. Many careers that require years of specialist training, such as yourselves would want to do, are now within the reach of AI software, especially when combined with advancing robotics. These growing abilities of AI are due in part to what is called ‘machine learning’, or ‘deep learning’. This type of software was developed by IBM for its ‘Watson’ AI to play and win the US television game Jeopardy. It is now being applied in many ways.
- Deep learning software works by scanning large amounts of ‘big data’ and making statistical judgements about the best response, within the given parameters. It is designed to learn from its own performance, improving its responses as it processes ever-increasing amounts of data at ever-increasing speeds. Such programs can thus scan whole bodies of health or legal data for example, in order to deliver the most relevant and accurate answer to a question. They can then perfect their responses until the task has been mastered.
- This process is modelled on similar complex feedback mechanisms in the brain itself, or neural networks. These are far more complex than the AI models of course and have only themselves become better understood recently, due to significant advances in neuro-biology. These in turn have been driven by improved laser, imaging and other technologies. This example shows how improvements in different fields are highly inter-dependent, with progress in one area helping to unlock the potential in others. ‘Moore’s Law’, which states that computing power doubles every two years, now applies in a number of fields.
- Deep learning has already contributed to major advances in the area of so-called natural language abilities. Until very recently, AI had not been able to master some language skills effectively. But deep learning and the use of neural networks have allowed rapid progress in voice and language recognition and interpretation.[5] Accurate and instantaneous translation is now possible in multiple languages.[6] This advance will in turn boost the development of programs able to recognize and respond to human enquiries, and ultimately, to conduct conversations. This is a genuinely disruptive advance in the communication abilities of machines and humans.[7] Language barriers will literally soon become something of the past.
- This growing mastery of natural language has greatly increased the range of jobs that AI can now do, to include many roles that require interaction with people. This is in addition of course to the many functions that involve working with data. As one more pessimistic futurist put it, ‘software will eventually… invade every workplace and swallow up nearly any white collar job that involves sitting in front of a computer manipulating information’.[8] Driven by economic imperatives, more and more tasks will be condensed into a computer program and outsourced to an AI-powered machine.
- This will affect a huge number of jobs across all functions and sectors. This includes many STEM fields from computer technology itself to laboratory-based scientific work. Deep learning is itself a form of pure science, with its careful repetition and rigorous analysis of volumes of data according to pre-defined parameters and objectives. One groundbreaking AI program, named Eureqa, was described by its creator as being able to figure something out in a few hours, that could previously have taken a scientist’s whole career.[9] The medical sector will also be deeply affected, as human tasks are re-assigned to machines. AI and robots already match or out-perform humans in many areas from surgery to diagnostics.
- These technologies will also create new opportunities of course. This includes the many programmers and coders required to design, set up and operate AI software. Deep learning software is highly labour-intensive. Completely new roles will also be created, such as what might be called ‘medical engineers’ for example, to put together the complex parts of a bionic eye or limb. But obsolescence is also built in. Ratios of 1 technician to 20,000 machines from a few years ago are already out of date. The billion-dollar data centers that power all this technology are operated by tiny workforces.[10] Humans are currently still required to direct and oversee AI and the machines on which it runs, but this could also change.
- Vast numbers of blue collar, service sector and more physical work functions will also be replaced, by a combination of AI and increasingly advanced robots. This is a continuation of much longer-term processes of mechanization, dating back to the industrial revolution. But this area is also experiencing dramatic transformations, as long-standing challenges are finally being overcome. Some human movements that had proved very difficult to replicate, such as the fine motor skills necessary to pick up a grape, are now being performed by artificial limbs.[11] This will allow new leaps forward to be made, with implications for all jobs involving repetitive physical tasks. This again will be experienced across the whole range of functions and sectors.
- Some of the advances in this area have been driven forward by a remarkable fusion of scientific, medical, and humanitarian impulses. This can be seen in the bionic, AI-powered limbs that have been developed by scientists who are themselves amputees and Para-Olympians.[12] Deep learning programs provide the mechanism that connects these artificial limbs to the human brain. This modelling of neural networks again relies on our growing understanding of neuro-biology.
- Advances in other areas of robot engineering have also contributed, themselves driven by rapid progress in different fields from lasers to nano-technology to batteries. The combination of advances in different fields in this way adds up to far more than the sum of its parts, with progress in one area greatly boosting the possibilities in others, as mentioned. This process is highly unpredictable however.
- A good example of how disruptive such advances may be for the workplace, is provided by the transport sector. This example also highlights how difficult it is to weigh up the likely positive and negative impacts. The sector has recently experienced accelerating advances in a number of areas, which cannot but lead to radical change. On the one hand are self-driving cars, which combine advanced engineering with the deep learning abilities of AI. On the other hand are advances in electric and other alternative-fuel vehicles, which have themselves benefited from recent progress in areas from batteries to materials.
- These advances promise, or perhaps threaten, depending on your viewpoint, to transform the transport sector beyond recognition. The changes are so radical that some industry insiders believe that the car industry in its current form will cease to exist in as little as two decades.[13] Millions of jobs worldwide are likely to disappear, in car manufacturing itself and the many related activities.
But the potential gains are also huge – from the expected greatly increased convenience and affordability of transport, to a reduction in the use of fossil fuels. Even the way that cities look and feel will change significantly, due to the space that will be freed up once cars are no longer owned en masse. Instead of being stuck in traffic jams, we may get around in super-fast ‘hyperloops’, as are already being planned by visionary entrepreneur Elon Musk and others.[14]
- Some believe that the benefits from the coming transformation will be experienced most deeply in these more indirect and unpredictable ways. Futurists such as Brynjolfsson and McAfee from MIT believe that the current advances in AI and other areas may provide as great a mental boost to humanity, as the physical one we gained from the inventions of the industrial revolution.[15] They also view a move away from current employment patterns as something fundamentally liberating for humankind. The wealth generated as the production frontier is pushed out by technological change to ‘vast and unprecedented levels’,[16] can then be used to support the pursuit of higher goals, as people no longer need to work for a living.
- Such an outcome depends on effective control and management of the new technologies and the wealth they will create, to ensure this is used for the greater good rather than contributing to more inequality. This is why policy-making, regulation, and education, especially in STEM fields, are all so important. The choices we make now in these areas can help shape and determine how these processes play out. This requires far greater public and policy discussion of all the relevant issues, including the more challenging aspects related to control and regulation.[17]
- lt is difficult to strategize about an unknown future, but there must now be far greater public analysis and debate about different future scenarios and their implications. There are a growing number of initiatives in this area of ‘future-watching’, as supply responds to demand, and such efforts must now be scaled up and magnified. Substantial investment in education in STEM fields must also become an urgent priority, in order to ensure that this analysis and debate is underpinned by well-informed and critical approaches.
Ladies and Gentlemen:
- As we look forward into an uncertain future, we can also learn lessons from history. The story of our recent past is one of previously undreamt of human progress due to scientific and technological advance, however uneven and incomplete. Life expectancy has tripled since the industrial revolution, while the lives of hundreds of millions have been significantly improved even in the past few decades.
- This longer historical perspective demonstrates just how far we have come. It also highlights the central role that has been played by political and social institutions in ensuring that the benefits have been spread as widely as they have. As scientist and historian Steven Pinker puts it, these ‘gifts of progress, are the result of the institutions and norms entrenched in the last 2 centuries’. These institutions and norms include reason, science, education, regulation, democracy and human rights. [18]
- We must now ensure that the even more amazing gains that are likely to be made this century are also managed effectively and for the greater good. This is possible through these same means – the insitutions and norms of our modern world. It is therefore incumbent upon all of us, whether as scientists, academics, educators, or economists, to work to strengthen further these institutions and norms. In this way they will be able to continue to play their vital roles in regulating and channeling change into a positive direction.
[1] Steven Kotler, ‘Tomorrowland: Our Journey from Science Fiction to Science Fact’ (2015)
[2] Eric Brynjolfsson and Andrew McAffe ‘The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies’, (2014)
[3] Martin Ford, ‘The Rise of the Robots: Technology and the Threat of Mass Unemployment’, (2015)
[4] Oxford Business Group (2013)
[5] https://www.theregister.co.uk/2016/12/09/improving_computers_learning_speech/
[6] https://www.theregister.co.uk/2017/07/17/tony_robinson_speechmatics/?page=2
[7] Ford, (2015) p.92
[8] Ibid. p.108
[9] Ibid., p.111
[10] Ibid.
[11] Kotler (2015), pp. 18-20
[12] Kotler (2015), pp.1-21
[13] ‘Mercedes CEO describes Inventions of the Future’, July 2017, http://tinyurl.com/yc2563cn
[14] https://www.wired.com/story/great-elon-musk-building-hyperloop/
[15] Brynjolfsson and McAfee (2015), p. 8
[16] Brynjolfsson and McAfee (2015)
[17] Ibid.
[18] Steven Pinker, in John Brockman (ed) ‘Know This’ (2015), p.