In this section, we will review the actionable items to mitigate the risks and leverage on the opportunities brought about by the A2 economy. The underpinning assumptions to these proposed solutions are multi-fold. Firstly, we are able to embrace A2 Economy as a society, in the spirit that those who can succeed will help those who can’t. Secondly, we have sufficient resources to make the jump into the A2 Economy given that there will be investments needed to bridge the initial gaps — business models, skills, legislature, social contracts, educational system etc.
So, “What is the A2 economy?” you may ask. That is a great question. To be frank, there is no clear answer or definition until we get there. However, it will be characterised by 4 Ps:
The 4Ps Model of the Artificial Intelligence and Automation-based Economy (Choy, 2019)
More specifically, we can expect:
1. AI and automation permeating into every part of culture, society and human psychology and part of this will involve biological and social integration of machines with humans. For social integration, we may have robots as our colleagues and bosses. With biological integration, we may have intelligent robotic arms or body parts. The implication is that once the integration ramps up, it will be extremely difficult for humans to be dissociated from machines in the future. The creation will now be part of the creator.
2. Algorithms underpinning automation at the start with AI taking over once there is sufficient data (training) to decipher intents and drive future machine behaviour. Just like how a young person can pick up good and bad habits, this ‘training’ needs to be carefully monitored with AI trainers being vetted for ethical and moral standing. A case in point was the launch of a chatbot a couple of years ago and the users bombarded it with colourful language, resulting in it learning and using the language with the other users interacting with it. Once the AI is trained, it will be difficult to ‘untrain’ it — just like getting someone to stop using coarse language or smoking.
3. Layering on of standards entrenches the technology globally. It is likely that once any AI and automation standard has been established globally, it will be difficult to switch to a new system. The legacy will be global and long-term. Hence, it also means that the owner of the standards (if it is not open source) will have a huge advantage over others. Singapore needs to own part of these technical and social construct standards, in order to establish influence globally and over its populace. For example, with regard to social construct standards, difficult conversations need to be set up to determine the machine-people divide (e.g. Should an AI doctor inform a patient that the patient is dying? Can people legally marry their cyber girl/boy friends? Can robots inherit wealth? What makes a humanoid or cyborg human or not? Even though many of these questions may not be relevant now, if Singapore wants to establish influence over these issues, we need to lead the field in conversing about them now before the technological disruptions take place in 5 to 10 years when chronic global unemployment sets in.
4. The strength and projected power of a nation won’t be about the physical size of the country or its population size — it will be about technological superiority. Singapore, despite its physical limitations, will have a real chance at establishing global superiority and punching above its weight (and size). Augmentation with AI will be key. How fast we move into AI and leverage on its strengths can give us a tremendous first-mover advantage.
So, where do these projections leave us? What can we do to get into the A2 economy as painlessly and quickly as possible?
5 Recommendations to Implement for the A2 Economy
Listed below are 5 areas that we need to work at if we wish to embrace the A2 economy as a society:
1. Create Avenues for AI Augmentation with Human Expertise
It is not absolutely clear how AI-driven robots will work with humans in the future. There are many possible modes and degrees of human augmentation (Thomas, 2019). Whether machines augment humans or vice versa and to what degree will be areas for discussion.
Two key factors are at play here:
a) Degree of Autonomy (who/what controls the relationship and actions)
b) Initiating Agent (who/what sets off the chain of responses)
Let’s examine a few of these possibilities.
· Level 1: Machines extending the expertise of humans
Intelligent machines can be trained by practitioners and experts to augment their work. For example, AI can be trained to support professionals to work i.e. a consultant, designer or marketer trains the robot to perform the same set of tasks to achieve superior results. For example, a master chef trains the restaurant robot to use his recipes to cook or a surgeon uses robotic arms and visual recognition cues to make decisions on how and where to make surgical incisions. When we consider this structure and mode of augmentation, the skillset of the expert suddenly becomes quite scalable. At level 1, the human expert is still the initiator and owner of the augmentation process. The expert will still maintain control over how and when the robots train other humans or robots. There is likely to be monitoring of the results from this level of augmentation whereby machines extend the expertise of humans.
· Level 2: Machines augmenting human performance in an autonomous manner
Currently, AI to check radiography scans is a popular example of how intelligent machines are augmenting radiographers’ work in a scalable manner. Often, the accuracy of these scans conducted by AI can be much higher compared to expert radiographers. The scans are also much faster which can be critical in certain medical cases when the patient may have late stage cancer, for example. The other example often cited would be AI conducting research on past legal cases, processing and indexing information for the lawyers and judges, thereby reducing time and resources required to fight a case for the client. Both examples are pegged at level 2 of human augmentation because the initiating agent is human but some level of machine autonomy is provided for with machines making baseline decisions to highlight or surface cases based on machine learning or algorithm.
· Level 3: Humans augmenting machine performance with some measure of human control
Here, the balance begins to tilt with the machines taking on more autonomy and initiating some of the work processes while humans take a back seat to the process. Here, autonomous vehicles allowed to navigate the roads with a human driver ready to take over at any point in time, especially in potentially difficult situations, could be an example of level 3 augmentation. One possible scenario put forth by proponents of self-driving technology is to have autonomous trucks drive between states on their own and then have a human driver hop in to take over when the truck is at the fringe of the city, to complete the last mile where the denser human population and traffic could pose some difficulties to the technology.
· Level 4: Humans augmenting the intelligence and capability of machines
At this level, intelligent machines are in the driver’s seat and will control the process along with giving instructions to humans to regulate their behaviour. For example, the Cogito emotional intelligence software (https://www.cogitocorp.com/solutions/) directs call centre operators to modulate their tone, use emotive phrases in order to increase empathy, provide a better pace and subsequently, engage better with the customer on the other end of the telephone line. With Cogito’s feedback (in bar chart format) providing live update to the operator on the device in front of the human operator, the human will adjust his or her speech outputs according to the machine’s feedback. If the feedback is ignored or the indicators do not improve during the duration of the call, the data could be sent to the operator’s superior for follow-up action. In this regard, Cogito very much initiates the feedback, controls and regulates human behaviour to achieve the optimum performance. There may be little autonomy on the part of the human to move outside of the acceptable range of behaviour.
Based on this framework of human-machine augmentation, it will be useful to determine how the augmentation can be maximised and the risks mitigated for each level so that ethical issues are addressed early on. Drawing up some rules and laws for humans and machines to observe may also be useful especially in the training of humans so that expectations are managed. For example, with robots as colleagues or personal assistants, humans need to manage their emotions if these colleagues perform better at certain tasks than them. Hence, a certain level of resilience and robotic (instead of emotional) quotient is required for humans to work with robots.
This also means that we need to review the current legislature concerning technology-enabled work, especially in the light of how AI may augment human work and vice versa. There are scenarios where AI supervises worker’s performance by analysing human outputs (e.g. Amazon staff’s productivity levels in moving goods in and out of warehouse, Cogito giving feedback to call centre operators on quality of response to customers) or when humanoids and humans work closely together to achieve an outcome that clear boundaries need to be drawn, going back to the almost absurd but not impossible situation that humans and robots may one day be legally married or robots inheriting a human’s wealth.
Educating our young on the boundaries will also be critical to maintain the uniqueness and purity of the human race although this will become increasingly difficult to defend as technological advances will blur the human-machine divide in time to come. Here, we could be talking about 70 to 100 years into the future when human-machine cyborgs become prevalent.
2. Transform Singapore into a Global A2 Economy Hub
The goal here is to transform Singapore into a global hub to leverage on the A2 economy. What MNCs will need in the future are highly secure and networked regional and global centres to manage their workforce of robots and intelligent machines. The pooling together of these control centres of MNCs and large enterprises in Singapore requires massive digital infrastructure, watertight cybersecurity firewalls, strong legislature to protect IP and fight fraud, a highly educated digital workforce to manage and control these intelligent machines globally. It is possible that putting these AI-driven centres in close proximity can reduce time and barrier for the AI (belonging to different enterprises and entities) to communicate among themselves, to effect trade and services.
An A2 economy hub requires the setup of several ‘spokes’ to support the A2 economy hub. These spokes include:
1) Subject Matter Experts who are Digital Natives
This will require the development of new subject matter experts (SMEs) who can innovate and manage intelligent machines. These SMEs should be able to pre-empt issues and spot developing errors based on data and trends. A high level of mindfulness concerning the environment, operational contexts and process changes will be needed to perform this task well. Having the courage and clarity of mind to intervene when systems are seemingly going well will also be critical for these SMEs. They will have to operate in environments where there is ambiguity and few rules.
2) Home-Grown MNCs and enculturated with the Country’s DNA
Each country will also have to grow her own pool of large MNCs so that she can exert sufficient influence on global standards and systems e.g. financial, technical and educational. A case in point is FaceBook launching its ‘Libra’ cryptocurrency, which disrupts the current US Dollar-denominated financial system. These attempts to bring about new global systems signal a new meta-national mentality (beyond MNCs pushing for globalised trade). It portends the intent of these global enterprises to grab mindshare and influence culture at the individual level. While this is nothing new (think ‘McDonalds’ and ‘Coca Cola’), the advent of mobile technology has given these enterprises a huge leverage as the influence occurs at the personalised level in high frequency (with users checking FaceBook, Twitter, WeChat or Whatsapp every few minutes). Online shopping (Amazon, Alibaba) and entertainment (NetFlix and YouTube) are other examples of enterprises engaging with their customers on possibly a daily basis. This implies deep globalised influence on the individual. Any governmental attempt to ban Netflix or FaceBook, for whatever reasons would trigger an uproar from their electorate. However, with deepfake technology (e.g. creating AI-enabled fake videos) and wayward social media influencers, governments will need to reconsider how to strengthen the social pact they have with the people they are governing. MNCs enculturated with a certain way of doing things could influence the global AI space over time.p3) Legislature to Resolve AI-associated Disputes
A legal framework needs to be in place to guide the resolution of AI-related issues. The framework should increase transparency and not place too much onus on AI designers to bear the brunt of the responsibility beyond what is expected, if and when their AI products and services do not perform to the level expected or result in losses to either property or lives. A balanced legislation will spur innovation while providing sufficient protection for claimants.
4) A Registry of New Innovations
Each country needs a registry of new innovations (beyond the ambit of the patent office) to catalyse partnerships and synergise research and development. The new strategic digital office can provide leadership to enterprises, pilot these new technologies and spur growth in identified areas.
3. Green Lane the Development of the Next Generation of Experts
While this is controversial, it may the only way to ensure that humans will continue to drive innovation within the respective industries. This implies early identification of promising workers and providing specialised training so that they move up the skills ladder from novice to expert, in an accelerated fashion. Experience is gained through error-based learning pedagogies while they are given learning stipend when acquiring baseline competencies despite machines performing these tasks at levels at or above baseline competencies (e.g. book-keeping).
Protecting baseline work for this group of identified young practitioners is critical so that they work through the competencies to gain the experience. Without this protection, human workers will never have the opportunity to perform these baseline work tasks and as such, may never be able to rise to become an expert in the respective fields.
Hence, at a young age (perhaps, 30 years old), these fast-tracked, hot-housed practitioners will have to gain sufficient expertise and experience to be considered top in their industry. They will begin managing the robots and pre-empting possible errors and issues. The other critical roles they will perform include conducting research and developing innovations.
4. Make School-Work Transition Seamless
The current school-work divide is a recent phenomenon as children in the past (pre-1950s) often work in farms or help out with work as part of their induction process into adulthood. When our economy developed, education then became a universal right for children to be given protected time to engage in learning. This divide results in children being exposed to authentic work only in their young adulthood (e.g. in their 20s, during internship during university), leaving it very late for them to develop a growth mindset, correct work ethic and psychological resilience. In the future, these attributes will need to be developed earlier, as early as during primary education so that our workforce can maintain their goodness of fit with the A2 economy.
The question is how we can make the school-work transition seamless. Described below are some recommendations for consideration:
· Merge learning and work where children accumulate ‘credits’ as they start ‘work’ at primary school and adults are rewarded to achieve new learning through challenge problems and makerspace programmes so that adults never stop learning while children start work in school, blurring the school-work divide. Adults who learn are given credits that can be used to further their learning and/or other basic needs (e.g. pay for electricity and water bills).
· Further develop ‘maker space’ as an educational or training pathway so that people learn as they create products. To be clear, these products are not limited to physical objects (e.g. tables or glassware) but includes digital products (e.g. codes, websites, apps and curriculum). Maker space education as a training pathway could be fuelled by enterprises looking for potential talents while at the same time, putting in some investments to trial new ideas and innovations. If there are talented ‘makers’ present, enterprises may consider the maker space as an alternative to their rapid prototyping or development arm.
By recognising ‘maker space’ as a training pathway, we can give new meaning to learning and innovation. In the new A2 economy, this form of training can be very useful for increasing the stream of new innovative products, giving meaning to people as they drive their own learning and subsequently, creating their own jobs. Instead of subsidising training places in formal training programmes, some form of stipends could be provided as people learn and create useful customised deliverables for enterprises. Either online courses or mentors (humans or AI) can be utilised to provide support to these ‘makers’ as they work through their products over 3 to 6 months in a physical community (akin to co-working space) which provides psychological and skill support. For makers, the portfolio-building exercise along with their interaction with enterprises which could become their future employers, are valuable assets. To this end, maker space education could be a viable albeit fluid form of education. Government subsidies, if any, will have to be outcome-based, only disbursed upon on delivery of agreed products by the makers. Documentation and accountability measures will be different from current CET training pathways and these can be drawn from our experience with internship and workplace learning initiatives, for example.
· Conduct competency profiling and provide learning recommendations throughout the individual’s academic and working life where learning is ‘pushed’ out to individuals to achieve goals and to be ‘badged’ or given tax exemptions to supplement course subsidies. These profiling and course recommendations processes can be undertaken by AI.
· Build ‘adaptive’ curriculum which allows for the emergence of ‘stem-cell’ workers who can be parachuted into any industry quickly and then move into new industries with minimal retraining. These workers will have to be imbibed with adaptive and growth mindset, equipped with deep generic skills (e.g. research, innovation, critical thinking), STEAM (Science, Technology, Engineering, Arts, Mathematics) competencies and can be mobilised to do research and drive new innovations whenever a newly emerging industry or field is discovered.
In short, learning and work should be viewed as an integral part of many cycles that occur in a person’s life with both activities rebalancing constantly as one’s age and learning needs change. The focus is always on improving the person’s work performance and standing in the society.
5. Expand Work Structures to Include Learning and Innovation
Going forward, current work needs to be reimagined, redefined and expanded. I do expect work for humans to fall into these three general categories in the future A2 economy:
A) work that still needs to be performed by humans
B) work to provide recreational value and entertainment
C) work to sustain and develop society
While not exhaustive, these 3 categories of work for humans could, in the future, define most of the work available for humans to perform in the world we live in. The remaining proportion of work could involve illegal and immoral work types which we will not discuss here although this 4th category could grow if we do not manage the A2 economy well.
A) Work that still needs to be performed by humans: Innovation and Expertise Work
Possible job titles: space exploration, expert, robot and people managers, social innovators, nanotechnology researchers, AI and robot trainers, Infrastructural system managers, strategists
It will be impossible to list all the possible jobs and sectors that humans will still have to perform in the future. The underlying principle is that these are jobs which require human innovation and risk-taking, to drive change based on pure human imagination and vision.
1. Green field technologies — many of the new and yet to be discovered technologies (i.e. the equivalent of Blockchain, nanotechnology in the past) will emerge quickly and may last a decade or so before AI takes over to drive the routine procedures. While we cannot identify them at this point in time, we will need industry planners to spot and grow winners to become unicorns in the future.
2. Robot Management — this includes the maintenance and operational requirements of robots although robots will eventually take over this job as well, in the distant future.
3. Environment — With the exponential increase in robot usage, we will have to review the recycling and disposal of robots and other hardware. Extracting rare metals and valuable components for reuse will be an industry that will be needed in the future.
4. Energy — The appetite for energy will be unceasing but hopefully, with innovation, we can find more efficient ways to utilise renewable energy and satisfy most of our energy needs in the future.
5. Space exploration — We will have an industry to explore space in the future as space transportation becomes more affordable and a reality for the masses, possibly in a few decades.
This category of work will involve people pushing boundaries to find new fields and discoveries. When augmented by AI, the discoveries and innovations in the future can be mind-boggling. It can be a really exciting future, if we manage our societies correctly.
B) Work to provide recreational value and entertainment
Possible job titles: athletes, craftsmen, artisans, literary writers and artists
With AI and automation taking precedence in many of the work processes in enterprises and governments, there is a need to balance the work mix in society so that humans still maintain a level of dignity and self-confidence driven by their work for other humans.
To a large extent, AI and automation can perform many of these tasks so some form of protection will be needed to allow for the transmission of human values and beliefs to our young and elderly. These sectors described below will be familiar to many of us as government retain these sectors for strategic and political reasons.
1. Recreation and Entertainment (e.g. sports, music and art) — There will be time for people to enjoy performances by humans.
2. Culture — Artisans and traditional craftsmen will still be needed to instil a sense of history into our young.
3. Fine dining — Great food will still be appreciated by humans and this could develop into a culture for foodies globally.
4. Human endurance activities — To challenge one’s limits, there could be endeavours to ‘break the record’ (e.g. fastest trip round the globe on foot)
5. Human-only communities — The government may wish to establish communities of people where they can engage in nature (e.g. gardening) without the presence of any machines. This is a form of ‘de-machination’ — just humans with humans so that people socialise with each other surrounded by nature, either for short retreats or long-term living.
6. Tourism — Being freed up to travel and enjoy the world will continue to be a draw for many people especially those who have the means to do so. Hence, developing innovative tourist attractions could be lucrative for the interested.
C) Work to sustain and develop society
Possible job titles: life coaches, prison wardens, life counsellors, family counsellors and programme coordinators, educators, social ethicists
An important part of the focus in the A2 economy will be to ensure humans who are not involved in innovative work (Category A) and recreational work (Category B) will still find meaning in what they do. Hence, Category C work is designed to:
· transmit human values and beliefs
· give humans a sense of self-worth and value
· ensure that humans still socialise and work together for good, especially of the larger community
· help humans maintain their physical, psychological and mental health
1. Life coaching — Students and young adults will need to learn how to work and learn together in small groups, picking up positive emotions and traits such as resilience and love, educational psychologists
2. Prison service — With humans, there will always be a need for punitive measures to correct behaviour. Hence, incarceration and rehabilitation work will likely continue for the foreseeable future. Prison wardens and counsellors need to review how they work with humans in the new context of A2 economy. There is a strong possibility that people who fall through the cracks will not have the option of getting back into the mainstream of society if the society does not reach out and structure the rehabilitation process.
3. Family services — Families will have to compete with intrusive technology (beyond television and mobile phones) to vie for each other’s attention. Family counsellors and family programme coordinators will need to step up their service offering if they wish to make a greater impact on family cohesiveness. The extent technology is allowed into the home will be an issue that families have to address now.
4. Education and expertise development — In line with the expectation that expertise will be critical for humans to maintain their value in society, educators will need to refocus their energies to instruct and guide their learners toward developing expertise rather than competencies only.
5. Social ethics — Ethical considerations leading to legislature will become even more critical in the future as machines push boundaries that one cannot even imagine right now. Hence, protecting human values and rights will be an important reason to keep a vibrant and strong social ethics community.
In essence, the proportion of work distributed across these 3 categories will have to be constantly rebalanced to maintain the well-being of the society and to keep enterprises at the cutting edge of innovation and productivity. Keeping the populace productively engaged through these different types of work will be a delicate balancing act that future governments have to grapple with. Not many governments will be able to do it well as fake news, social media and activist groups distort demands and needs. Engaging the society in sustained conversations will be critical so that social trust and respect are built as we step into an A2 enabled future.
This paper has considered somewhat extensively the following areas:
· Section 1: How will Work and Learning be Disrupted?
· Section 2: Peering into the AI-Enabled Future
· Section 3: What Can We Do About the A2 Economy?
We have tried to maintain a balanced perspective to the future of work and learning without being overly paranoid about machine domination and the demise of the human race, keeping a healthy dose of optimism about the future where humans can have quality time to rest, think and create. Still, there are challenges ahead with major disruptions to the current systems, as we know them. Notably, our system of work and remuneration, our educational system comprising essentially of pre-employment training, our social structure, our government and more importantly, what it means to be human will have to be redefined, sharpened and change to fit a brave new world in the A2 economy.
Michael Choy (Dr.)
01 Aug 2019
Join my LinkedIn Community here: www.linkedin.com/in/michael-choy-Dioworks