Governance Challenges in the Age of Augmented Intelligence (AI): The Funny Business of Disruptions
P.K. Agarwal
Abstract
Beginning with the mainframe computers of the 1960s, we have witnessed wave after wave of tech-driven innovation. Interestingly each wave of technology becomes stronger than the previous one in its scope and reach causing disruption in its wake. Naturally, with each wave, some jobs are impacted but economic growth and new opportunities have so far compensated for these losses by the creation of new jobs. However, the newest wave propelled by Internet of Things, big data, machine learning, and artificial intelligence is exponentially stronger and has the potential of social and economic disruption at an unprecedented scale and speed.
There is a compelling need for public administrators to pay attention to these emerging forces and be prepared to smooth out the resulting social and economic impact for a large segment of society. The public administrators are likely to see a new set of challenges not witnessed before. There will be new demands that will strain the resources of government. There is an urgent need to rethink the role of government and lay the groundwork for serving the interests of the society.
Technological Disruption, Nothing New
By the 1870s, New Yorkers were taking over 100 million horsecar trips per year and by 1880 there were at least 150,000 horses in the city. Some of these provided transportation for people while others served to move freight from trains into and around the growing metropolis. At a rate of 22 pounds per horse per day, equine manure added up to millions of pounds each day and over 100,000 tons per year (not to mention around 10 million gallons of urine). Per one observer at the time, the streets were “literally carpeted with a warm, brown matting . . . smelling to heaven.” So-called “crossing sweepers” would offer their services to pedestrians, clearing out paths for walking, but when it rained, the streets turned to muck. And when it was dry, wind whipped up the manure dust and choked the citizenry.
In the late 1800s, the “manure crisis” had become a problem in every major city in the world. The Times of London in 1894 did a linear extrapolation leading to a forecast that by the 1950s, London would be buried under nine feet of manure [1]. Without the benefit of history, one could easily romanticize the late 1800s as the pinnacle of clean technology. That was clearly not the case. This massive amount of horse excrement and carcasses were a major environmental and public health hazard.
A new disruptive technology was on the horizon. In 1870, Siegfried Marcus had built the first gasoline-powered combustion engine [2]. Karl Benz developed the first motorized vehicle in 1885 [3]. Henry Ford built his first experimental car in 1896 and formed the Ford Motor Company in 1903. By 1915, the Ford Motor Company was producing a million vehicles annually, and by 1927, it had produced over 15 million cars [4].
The disruption of the horse business was well on its way:
“Dispense with a horse and save the expense, care and anxiety of keeping it…”–An 1898 advertisement touting the benefits of owning a car (Figure 7.1).
Figure 7.1 Early automobiles
Cars were replacing horses in the cities and tractors, as an alternative to horses, had started to appear on American farms. In 1917, Henry Ford introduced the Fordson, a wildly popular mass-produced tractor that ended up capturing 77 percent of the market in the following six years [5]. Tractors became the norm on American farms. The disruption of the horse was in full force on the American farm. Thus began a sharp, long-term decline in the number of horses used on American farms and cities, and a corresponding exponential increase in the number of tractors and cars.
The number of horses in the United States peaked at around 26 million in 1915 and steadily declined to around 3 million by 1960. Life became increasingly difficult for those in the horse business, whether in the cities or on the farms. This was true not only for those in the horse business but also for the ones lacking skills relevant for the emerging new economy. Farm unemployment was starting to emerge as a major social and political issue. The First World War temporarily alleviated the job crisis but jobs became even a more serious issue at the end of the First World War as soldiers came home. Such is the nature of disruption that accompanies innovation.
One hundred years later, we are in the midst of a social and economic transformation that is eerily similar to the one brought forth by the internal combustion engine. Except this time around the disruptions are being caused by computer technologies sweeping our landscape. During the Industrial Revolution, we saw the horses and those in the horse business lose their jobs. In the current disruptive cycle, any job that is predictable and repetitive is at risk. Many call this the Fourth Industrial Revolution or Industry 4.0. Fast emerging technologies, such as Artificial Intelligence, Robotics, Machine Learning, Internet of Things, Blockchain, and Big Data propose to transform the economic and social landscape in the next couple of decades. These technologies will affect a very significant number of jobs in every country in the world. As if the Industrial Revolution was not disruptive enough, McKinsey Global Institute commented on the current situation: “We estimate that this change is happening ten times faster and at 300 times the scale, or roughly 3,000 times the impact” [6]. A forthcoming change of such intensity is going to create a very serious challenge for government. Moreover, it is currently not on the radar of most governments across the world. If governments do not begin to address this issue now, there is considerable risk of being late to the party and face the wrath of the large number of people impacted (Figure 7.2).
Figure 7.2 Early 20th century automation trends
Disruption Comes in Waves Upon Waves
One of the key ideas in Steven Johnson’s book, Where Good Ideas Come From, is that evolution and innovation usually happen in the realm of the adjacent possible, that is, innovation takes place adjacent to something else that is already in place. He gives the example of eBay, which could only happen until (a) someone had invented computers, which led to (b) a way to connect those computers, which led to (c) smaller computers, and then to the World Wide Web, and then eventually to eBay as a platform for online payments. Similarly, Uber could not have happened without mobile phones and the GPS. The notion of lone genius is a rarity. This “adjacent possible” creates a series of sequential waves and generally each wave tends to be stronger than the previous one in its scope and reach and its disruptive potential.
While electronic digital computers were around since the 1930s, the commercialization began with the mainframes in the 1960s. That led to wave after wave of innovation in the hardware front. The mainframe wave was followed by mini-computers, then PCs and networking, and then to the mobile devices. Similar innovation was taking place in the world of software. The convergence and maturity of hardware, software, and networking gave us the Internet. That led to the waves of application of the Internet. The first wave was the Internet of documents, followed by the Internet of commerce, the Internet of people (social media), and more recently the Internet of Things (IoT). Each wave builds upon the momentum created by the prior one and carries greater disruptive impact.
To clarify the increasing impact of each wave, Internet of documents gave us web pages and e-mail, initially connecting a few hundred million people. The next wave of e-commerce multiplied the user base by a factor of three to five. The next wave of social media or the Internet of People connected half the humanity. The newest wave of the Internet of Things will connect 20 to 30 billion things over the next four to five years [7]. The IoT will create a new world in which all things are connected and talking to each other and us.
These 20 to 30 billion connected things will generate massive amounts of data. Add to this to the data resulting from the innovations of other waves; we now face “infinite data” far beyond the ability of any human being to comprehend. This brings the tools of Big Data, Machine Learning, and AI to help make sense of this data. In the AI circles, there is considerable debate whether the rise of the machine is Intelligence Augmentation (IA) or AI. I believe it will be both. In some situations, machines will enhance and support human decisions and in other cases replace human in the performance of tasks. The AI or IA machines will need large quantities of data to learn the business. Accordingly, Clive Humby coined the phrase “Data is the new oil,” an apt metaphor for the relationship of data to learning machines.
Disruptions Redistribute Income and Create Income Disparities
It is hard to comprehend the extent of new wealth creation that happens alongside with disruptions. Imagine taking the valuation of FAANG (Facebook, Apple, Amazon, Netflix, and Google) companies and shop for established companies. The current combined valuation of FAANG companies is in excess of $3 trillion. Let us start with pricing the five most valuable automotive manufacturers: Toyota, Volkswagen, Daimler, BMW, and Honda. Their collective market cap is approximately $400 billion. The total market cap of five largest airlines— American, Delta, United, Emirates and Southwest—is about $31 billion. The five most valuable fast food brands—McDonalds, Starbucks, Subway, KFC, and Dominos—total to $210 billion. The world’s largest retailer, Walmart, has a market cap of around $300 billion. The point of this is that the valuation of all of these well-known brands put together account for less than one-third of the valuation of the FAANG companies. I hope this provides a glimpse of the extent of new wealth creation. Why? Entrepreneurs upset the status quo and create new opportunities. These new disruptive companies show enormous potential for growth, both by from taking market shares away from existing players as well as creating new markets that did not exist before. Naturally, this new wealth tends to be concentrated among a few; entrepreneurs willing to take risks and those with skills that support the disruptive industries and ventures. In part, this would help explain the income gap and the shrinking middle-class phenomenon that we currently face. This has happened before. We saw the emergence of robber barons in the gilded age during the Industrial Revolution. To give an idea of the size and scale of this new wealth creation, the market cap of tech or tech-driven companies that have gone public in the last 35 years is in the $15 to $20 trillion range. By comparison, all of the gold in the world is valued at $8 trillion; the size of the U.S. economy is approximately $20 trillion [8]. At the individual level, of the top ten richest men in the United States (sometimes referred to as the three-comma club since a billion has three commas), seven have made their wealth in the tech or information sector. The collective net worth of these seven individuals is in excess of half a trillion dollars, equivalent to the entire economy of Sweden.
The Impact of Disruptions on Jobs
While the “horsepower” fueled the disruption during the Industrial Revolution, the current waves of disruption are fueled by computing technologies and its latest manifestations in the form of Machine Learning/AI and Robotics. These technologies are affecting the blue-collar as well as white-collar jobs. A recent report by the World Economic forum provides an interesting perspective on the job churn through 2022 [9].
One set of estimates indicates that 75 million jobs may be displaced by a shift in the division of labor between humans and machines, while 133 million new roles may emerge that are more adapted to the new division of labor between humans, machines and algorithms. While these estimates and the assumptions behind them should be treated with caution, not least because they represent a subset of employment globally, they are useful in highlighting the types of adaptation strategies that must be put in place to facilitate the transition of the workforce to the new world.
As computers become more powerful and software more adept, we will see computers being able to mimic the five human senses. After all, it is our senses that define the humans to the rest of the world. With our computers interacting with us along the five senses, the distinction between humans and machines would begin to diminish. AI and Machine Learning will create human-like cognitive capabilities in computer hardware and software. The current applications of Machine Learning include natural language processing, handwriting recognition, computer vision, and other sensory perceptions. The keyboards will disappear as an artifact of the past. Keyboards were only necessary to provide very precise instructions to a slow, dumb computer. That is no longer the case and therefore a steady phase out of the keyboard. Further, what sets AI apart from prior technology waves is its ability to learn, then evolve, and take on more complex tasks and decisions. Thus, one can imagine that the more predictable and repeatable a job, the greater its potential of being taken over by machines.
A combination of Machine Learning and Robotics will replace low-end, labor-intensive jobs such as picking fruits and vegetables, cooking food, or car washing due to their predictability and repeatability. A recent startup in San Francisco, CafeX, now has a robot serving as a barista. There is early signs of the encroachment of machine learning (bots) into the jobs such as financial and sports reporting and financial advising. Machine Learning systems are showing great promise in interpreting images in the field of radiology and pathology.
Probably the most spectacular product of the Machine Learning and AI revolution is the autonomous vehicle. The potential for job disruption by the autonomous vehicle is huge. For starters, the transportation sector accounts for 7 percent of the U.S. economy. Approximately 5 million people are in the driving profession, which includes truck drivers, bus drivers, transportation for hire etc. A recent NPR report pointed out that truck, delivery, and tractor driving is the most common occupation in 29 of the 50 U.S. states [10]. The autonomous vehicles will also change the service model of personal transportation from ownership to transportation as a service, thus reducing the demand for passenger cars. This will affect jobs in areas such as vehicle sales, insurance, mechanics and engineers, parking lots, retail, and fuel. The health care sector will also feel the brunt of autonomous vehicles as the number of auto accidents would reduce dramatically (90 percent of all accidents are due to human error). The annual economic cost of road traffic accidents in the United States alone is around $242 billion (1.6 percent of our 2010 GDP). All this translates into jobs [11].
New Jobs Displace Old Ones
There are two distinct schools of thought of the future of jobs. The first group (pessimists) believes that we are destined for a jobless future in which machines will take over most of the jobs leading to a social crisis and anarchy unless public administrators act now. The other group (optimists) believe that disruptions create new wealth, new products and services, and new consumers. Economy will grow and standard of living will go up. However, both camps agree on one thing—many current jobs categories will disappear or shrink while new types of jobs will be created. Schumpeter’s creative destruction in action! According to an Oxford study, 47 percent of all jobs are at some level of risk due to disruptions driven by cognitive capabilities of computers [12]. Therefore, no matter what, there is an upcoming massive need for reskilling people.
I am on the optimist side of this debate. Over the course of time, people become more productive due to disruptive technologies and economies grow. We create new products and services that could not have been even imagined a decade or two earlier. The more nimble of the existing companies transform themselves while new companies and industries emerge. Many countries, such a Singapore, reposition themselves to be successful in the new economy. Others go into a decline. Accordingly, we are starting to see the emergence of new job classes. Within the last ten years, we have witnessed the creation of new jobs that never existed before. Here are ten new job types that are growing very rapidly.
For instance, the number of app developers worldwide is estimated at eight to ten million. The number of registered drone pilots in the United States now exceeds 100,000 and growing. The U.S. Bureau of Labor Statistics forecast a 29 percent annual growth rate for genetic counselors between 2014 and 2024. Our greatest challenge in the next 10 to 20 years is not about having enough jobs, but finding the right fit between jobs and the people, especially those people whose jobs have been displaced by disruption and unable to reskill themselves for this new world order.
|
1. App Developers |
|
2. Social Media Manager/Digital Marketing |
|
3. Cloud Computing Services |
|
4. UX Design |
|
5. Sustainability Expert |
|
6. Data Mining/Big Data Analysts |
|
7. Advanced Manufacturing Specialist |
|
8. Education/Admissions Consultants |
|
9. Genetic Counselor |
|
10. Drone Operator |
The Challenge for Governments
The AI-driven disruptive forces are also likely to cause significant change from the status quo for governments. Government will be forced into a significantly expanded role to manage the resulting crises.
Social Safety Nets
As AI and related technologies gain momentum, we are likely to see a large number of people whose jobs disappear and who would not be able to reskill themselves for the new type of jobs. Considering that robots would perform more of the routine manual work, public works programs such as the New Deal (or WPA) that provided relief during the great depression would have limited feasibility. To deal with the prospects of this potential mass unemployment coupled with aging demographics, many governments are experimenting with the concept of Universal Basic Income (UBI). A UBI is a type of program in which citizens would receive a regular sum of money from a source, typically government. UBI would have no means test and would be distributed to each person regardless of any changes to their financial status. There is no requirement to look for work and this would be independent of any other income. This is not a new idea and dates back to sixteenth century when Sir Thomas More argued this in Utopia. UBI is definitely gaining new ground in light of AI and has many proponents including notables, such as Elon Musk and Ray Kurzweil. Among the U.S. population, 48 percent of the people support this idea.
A more radical viewpoint on this issue of “survival wage” is to think of it as an income distribution problem. There is nothing wrong in principle if machines take over all of the mundane and repetitive work leading to abundance. The larger question is that who owns the means of production and for whose benefit. Speaking economics, in a scenario where machines do all the work, we could control the growth leading to an increased gross domestic product (GDP). In this scenario in which pie is larger and the population is more or less the same, could everyone not have a larger slice of the pie? In essence, if we could find a way to share this abundance, the entire society would be well off. In this scenario, it becomes a question of the collective political will of distributing income differently. Another less radical solution is to reduce the workweek from its current level in the United States of 40 hours a week. In the mid-1800s the average workweek was 70 hours. The workweek in the United States declined steadily since until the late 1950s when it averaged 40 hours. Since then it has remained at 40 hours per week. Henry Ford popularized the slogan of “ 8 hours for work, 8 hours for rest, 8 hours for what we will” in the 1920s. Is there a modern day Ford who is willing to change the “8 hours for work” to “4 hours of work”?
“Show Me the Money”
One of the more serious consequences of AI-driven disruption would be a change in the revenue streams of government. If machines are doing a great deal of work traditionally done by humans, who pays the taxes. That has led many, including Bill Gates, to suggest that we ought to tax the robots [13]. The EU had also considered legislation that would tax the robots and use that revenue to pay for worker retraining. That legislation did not go through but most likely come up again as the issue of worker retraining becomes paramount. On a fun note, last year, the Saudi Government granted citizenship to a robot name Sophia, made by Hanson Robotics [14]. This was a first of its kind—a robot to receive citizenship of a country. One would imagine that if a robot can be a citizen, the privilege of paying taxes comes along with it. Incidentally, last November Sophia was named as the United Nations Development Program’s Innovation Champion, the first nonhuman to be given a UN title.
Another example of how government revenues are likely to be impacted by the AI revolution is in the area of driverless cars.
A recent Deloitte Insights report states [15]: “The US public sector will likely have to figure out how to offset anticipated declines in the $251 billion annually generated from fuel taxes, public-transportation fees, tolls, vehicle sales taxes, municipal parking, and registration and licensing fees. All these revenues are tied to today’s reality of individually owned and operated vehicles—for instance, the need for parking diminishes with the rise of autonomous-drive shared mobility.”
It is hard to precisely quantify how much and how fast will the AI impact government revenues but some of the upcoming needs are obvious. The need for some sort of safety net, be it UBI or some other form, will rise soon. Investment in worker retraining on a large scale is a critical need. In addition, there is a critical need to upgrade the transportation, energy, and other urban infrastructures to support development of Smart Cities.
New Dimensions in Consumer Protection
Until recently, we had this utopian assumption that the AI machines will be free from human biases and therefore a progressive force for society. However, it is turning out that these machines are faithfully integrating the biases ingrained in the data used to train the machines.
John Giannandrea, Apple’s Chief of Machine Learning and AI Strategy remarked, “It’s important that we be transparent about the training data that we are using, and are looking for hidden biases in it, otherwise we are building biased systems. If someone is trying to sell you a black box system for medical decision support, and you don’t know how it works or what data was used to train it, then I wouldn’t trust it.”
Similarly, Tom Simonite, writing for Wired magazine and describing research by Professor Vicente Ordóñez, said: “Their results are illuminating. Two prominent research-image collections—including one supported by Microsoft and Facebook—display a predictable gender bias in their depiction of activities such as cooking and sports. Images of shopping and washing are linked to women, for example, while coaching and shooting are tied to men. Machine learning software trained on the datasets didn’t just mirror those biases, it amplified them. If a photo set generally associated women with cooking, software trained by studying those photos and their labels created an even stronger association.”
It is easy to postulate that all historical data will have some biases along the lines of gender, ethnicity, or other demographics. We now run the risk of institutionalizing and possibly amplifying these biases. These machines might even find other correlations that are not that obvious to us humans.
Another source of AI related bias is through algorithms. Algorithms are rules, provided by human beings or machines that assist AI machines in making decisions. With increased sophistication, these algorithms are becoming more complex and in many cases, no one fully understands them. There have been suggestions that we will need to design new AI machines to decipher the reasoning behind the decision making of other AI machines.
A 2017 report from AI Now Institute recommends [16]:
Core public agencies, such as those responsible for criminal justice, healthcare, welfare, and education (e.g., “high stakes” domains) should no longer use ‘black box’ AI and algorithmic systems. This includes the unreviewed or unvalidated use of pre-trained models, AI systems licensed from third party vendors, and algorithmic processes created in-house. The use of such systems by public agencies raises serious due process concerns, and at a minimum such systems should be available for public auditing, testing, and review, and subject to accountability standards.
eGovernance and Services
The AI tools could be a real boon for government services. For a significant portion of the user of government services, including the elderly, the access to technology and the learning of how to use it is still a challenge. While most of the government services are now available online, the usage of government online service other than tax filings and motor vehicles is still somewhat limited. AI-enabled services could be a boon by bridging the accessibility and tech literacy gap. The greatest potential lies in the use of voice recognition systems. Thanks to exponential improvements in natural language processing (NLP), machine learning-based voice recognition systems last year achieved an accuracy of 95 percent, a tipping point, putting it on par with humans [17]. In some cases, there are claims of 99 percent accuracy in low-noise environments. Voice-based systems are ready for mass consumer usage. The rapid adoption of smart assistants/speakers enables the infrastructure needed for the delivery of these services. In effect, voice is what the webpage was in the 90’s. It is the new browser.
While AI-enabled technologies offer significant opportunity, in general, governments are not well positioned to capitalize on these. They are barely playing catch up to the earlier generation of technologies. At present, more than half the traffic on the Internet is via mobile devices. Government are just getting around to migrating their services from the desktop to the mobile platforms. Now comes the challenge of voice enabling these eServices. Governments just do not possess these skillsets. There is an opportunity for the private sector to help bridge this gap. Done right, these could result in productive public–private partnerships. Done poorly, this could lead to de facto outsourcing of public services with government having little control over how these services get delivered.
On the innovation front, the EU believes that it should not matter where you live or how much you earn to have access to high speed Internet. To enable this, the EU is offering grants to localities to enable free Wi-Fi connectivity for citizens and visitors in public spaces such as parks, squares, public building, libraries, health centers, and museums everywhere in Europe through WiFi4EU.
Ensuring Individual Right to Be Left Alone
Louis Brandeis and Samuel Warren coauthored a landmark article for the Harvard Law Review in 1890 that defined protection of the private realm as the foundation of individual freedom in the modern age. This was in response to the new media of the time such as photography and other means of capturing and reproducing information. These new technologies gave increased capacity to government, press, and other entities too access previously unavailable details of personal activity. Brandeis and Warren argued for the law to evolve to deal with this technological change. Louis Brandeis subsequently as a member of the Supreme Court remained a champion of the “right to be let alone” as “the most comprehensive of rights, and the right most valued by civilized men.”
Now, a century later, new technologies are forcing a very similar debate over the individual’s right over their personal information and the right to privacy. The EU has taken a bold step with the General Data Protection Regulation (GDPR), which enables individuals to better control their personal data. This is still a work in progress and many assert that GDPR may be an overreaction to this growing problem. The Government of India is also considering similar measure to control the proliferation of personal information of its citizens and would prefer to keep it within the national boundaries.
It is the opinion of this author that data is now an asset as it has market value. Therefore, there is a critical need to establish who owns this data and what rights other related parties have over this data. At present, the de facto standard is that the various platforms own the data with limited control by the individual to whom the data pertains. For starters, one ought to have the right to opt out or better yet, to opt in for the capture of personal data. One could imagine a framework in which individuals own the data related to their personal, social, or commercial activity and they get to define the terms and conditions for the use of such data by other entities. This would make a significant impact in preserving privacy in the digital age.
Concluding Remarks
Governments are naturally slow to change because they have to be responsive to every single person in their constituency, including the proverbial “little old lady” who will go the DMV window to pay for her car registration in cash. Governments do not get to choose their customers. One cannot be like Amazon and choose the methods of delivery of services. On the other hand, technology offers new opportunities to reinvent government, and has the potential of making it more efficient and progressive. This gap is what keeps public administrators awake at night! When the pace of external change exceeds internal change, it can reach a tipping point and create a serious, sometimes irreversible, crisis for a government. With AI technologies, we are approaching that tipping point. There is an urgent need for public administrators to be proactive; they cannot simply wait. It is a rarity that even industry leaders are calling for the government to be more engaged, and more active in shaping policy and change. The time is now to seize the day!
References
Historic, U.K. 1894. “The Great Horse Manure Crisis of 1894.” Retrieved from https://historic-uk.com/HistoryUK/HistoryofBritain/Great-Horse-Manure-Crisis-of-1894/
ASME. “Siegfried Marcus Car.” Retrieved from https://asme.org/about-asme/who-we-are/engineering-history/landmarks/203-siegfried-marcus-car
ThoughtCo. “Biography of Karl Benz”. Retrieved from https://thoughtco.com/karl-benz-and-automobile-4077066
Ford Website. “Company timeline”. Retrieved from https://corporate.ford.com/history.html
Wikipedia. “Tractor.” Retrieved from https://en.wikipedia.org/wiki/Tractor
McKinsey and Co. “The Four Global Forces Breaking All the Trends.” Retrieved
from https://mckinsey.com/business-functions/strategy-and-corporate-finance/our-insights/the-four-global-forces-breaking-all-the-trends
Statista. “Internet of Things (IoT) connected devices installed base worldwide from 2015 to 2025”. Retrieved from https://statista.com/statistics/471264/iot-number-of-connected-devices-worldwide/
Wikipedia. “Gold Reserve” Retrieved from https://en.wikipedia.org/wiki/Gold_reserve
World Economic Forum. “The Future of Jobs Report 2018.” Retrieved from https://weforum.org/reports/the-future-of-jobs-report-2018
NPR Planet Money. “Map: The Most Common Job in Every State”. Retrieved from https://npr.org/sections/money/2015/02/05/382664837/map-the-most-
common-job-in-every-state
U.S. Dept. of Transportation. “The Economic and Societal Impact of Motor Vehicle Crashes.” Retrieved from https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/812013
The Economist. “A Study Finds Nearly Half of Jobs are Vulnerable to Automation.” Retrieved from https://economist.com/graphic-detail/2018/04/24/a-study-finds-nearly-half-of-jobs-are-vulnerable-to-automation
Quartz. “The robot that takes your job should pay taxes, says Bill Gates.” Retrieved from https://qz.com/911968/bill-gates-the-robot-that-takes-your-job-should-pay-taxes/
Wikipedia. “Sophia (robot)”. Retrieved from https://en.wikipedia.org/wiki/Sophia_(robot)
Deloitte Insights. “The Future of Mobility.” Retrieved from https://www2.deloitte.com/insights/us/en/focus/future-of-mobility/transportation-technology.html
Medium. “The 10 Top Recommendations for the AI Field in 2017.” Retrieved from https://medium.com/@AINowInstitute/the-10-top-recommendations-for-the-ai-field-in-2017-b3253624a7
Recode. “Google’s ability to understand language is nearly equivalent to humans.” Retrieved from https://recode.net/2017/5/31/15720118/google-understand-language-speech-equivalent-humans-code-conference-mary-meeker