“The best way to predict the future is to invent it.” – Alan Kay, a renowned computer scientist & ACM Turing Award WinnerIt was a warm July afternoon in the year 1825 when Richard Baxter (1) was asked to leave. His factory, like many others in the city of Manchester, had adopted the power loom which had reduced the demand for skilled handweavers, causing a decline in wages and widespread unemployment. Not much is known about Baxter following the key developments in the textile industry that led to his obsolescence – he presumably struggled to make ends meet for the rest of his life. As Keynes would later say, Baxter, like many others, was a victim of “technological unemployment” (2) which is but a temporary phase of maladjustment in an economy that embraces innovation.But with extensive automation fueled by the hyper-scale cloud, artificial intelligence, and advances in robotics, will the forthcoming revolution be even more disruptive? McKinsey Global Institute estimates that across the world, between 400 million and 800 million individuals could be displaced by automation and will need to find new jobs by 2030, based on various automation adoption scenarios. (3) To understand the magnitude of this problem, approximately 5-11% of the existing population of the world will need to find new jobs in the next 10-12 years. And to further elucidate the urgency of the situation, consider finance organizations that perform a wide range of activities from collecting basic data, accounting, cash management to making complex decisions and counseling business leaders. The potential for improving performance through automation varies across sub-functions and requires a portfolio of technologies to unlock the full opportunity. Applying the same methodology, McKinsey found that currently demonstrated technologies can fully automate 42% of finance activities and mostly automate a further 19% of the other tasks (4). Is this just the lull before a storm? For a storm is inevitable. ———————————————————In March, 2016 Google’s DeepMind AI AlphaGo outplayed Lee Se-dol, the eighteen time South Korean world champion of Go (an ancient complex strategy game played in a grid of 19 by 19), for the fourth time to win the overall series by 4-1. This gigantic defeat was powered the advanced systems built on deep neural networks and machine learning. (5) Is this the birth of a far more intelligent species? There is no doubt that Artificial Intelligence will become ubiquitous. Seeing its applications range from making driverless cars a possibility to eliminating mundane tasks of the industrial workforce, we fear a massive economic dislocation induced by robots. There is uncertainty whether this proliferation of Artificial Intelligence will ultimately be helpful or destructive to mankind. After all, if you were a radiologist who no longer has a job because a machine can read the X-Ray better or an unemployed lawyer who could not instantly spot legal patterns in millions of documents, you would be full of disdain and understandably so. Make no mistake, AI will drive large parts of our economies. However, we need to quickly move away from this myopic, dystopian world view and take charge. In his debut book, Hit Refresh, Microsoft’s CEO Satya Nadella envisages a future where collaborative robots or co-bots augment the abilities and experiences of humans, and together they can solve the greatest challenges of mankind – poverty, disease and ignorance. Fundamentally, we need to refocus on how human gifts like creativity and emotion can be mixed with powerful AI computation to help the society move forward. (6)To elucidate this view, take for example, surgical robots that enable administration of complex, minimally invasive procedures with better visualization, increased precision, and enhanced dexterity. These bots replicate the surgeon’s hand, wrist and finger movements into precisely engineered movements of miniaturized surgical instruments inside a patient’s body. In the past 14 years, over 1.75 million robotic procedures were performed in the United States across various surgical specialties and the risk of adverse events was as low as 0.0834%. Furthermore, smaller incisions, faster recovery time, lesser chance of infection and lower pain in the above procedures illustrate the powerful results that collaboration between humans and robots can achieve. (7)But, a relationship between man and machine can no longer be based on merely on Isaac Asimov’s Three Laws of Robotics introduced in his liberal science fiction writings. Symbiotic intelligence and partnership between bots and humans necessitate certain essentials from both ends. This brings us to the core requirements of designing Artificial Intelligence systems (a modest attempt to a profound philosophical and scientific problem, inspired by Satya Nadella) (5) 1. AI must be engineered to assist humanity – What is easily undermined, but strikingly difficult to achieve is to create a design framework which ensures all robots are designed with the right intentions. Researchers agree that there is no reason to expect AI to become intentionally benevolent or malevolent – its behavior is pre-dominantly determined by the motives and the imagination of the creators (8). To that extent, there are two likely scenarios in which AI might become a risk:a. AI is programmed to do something devastating: Consider autonomous weapons that are programmed to kill. In the hands of the wrong person, these weapons could easily cause mass casualties. Moreover, an AI arms race could inadvertently lead to an AI war that may in turn result in mass casualties. b. AI is programmed to do something beneficial, but it develops a destructive method for achieving its goal: If a super-intelligent system is tasked with an ambitious geoengineering project, it might wreak havoc on the environment as a side effect, viewing human attempts to stop it as a threat that it needs to overcome. We need to keep in mind that designing AI models is a recursive process based on multiple iterations of self-improvement. To fully understand the behavior, we need to test the model at varying levels of stress and initial conditions. 2. AI must be transparent – While robots will know things about humans, but it is also critical that we understand how bots see and analyze things around themselves. Transparency in design and awareness about technology will ensure that users can safely leverage AI for their benefit, without causing any harm to others. 3. AI must maximize efficiencies without harming the dignity of people.This is an outcome which is only possible when we engage broader and diverse populations in shaping the values and purposes inherent in AI design. Artificial intelligence must be based on representative research to strictly guard against socio-cultural biases, flawed heuristics and conscious or unconscious discrimination. 4. AI must be designed for intelligent privacy.Chatbots and virtual assistants like Siri or Alexa have deep knowledge about an individual’s idiosyncrasies apart from access to confidential information like bank-account passwords, personal files, photographs etc. Privacy needs to be built-in rather than a bolt-on to secure this information.5. AI should have algorithmic accountability. This is critical to ensure human beings can undo any unexpected outcomes, harms and consequences.——————————————————— By a twist of fate, in September 2012, Baxter was reborn – now as an industrial robot. He was built by Rethink Robotics, a Boston based company co-founded by the MIT professor and erstwhile Director of the Artificial Intelligence Laboratory, Rodney Brooks. (9) Contrary to what one may expect, Baxter did not wish to replace factory workers out of vengeance. Instead, he would work alongside humans helping them with simple industrial jobs such as loading, unloading, sorting, and handling of materials. The engineers, it seems, were cognizant of some of the above ideals. All this sounds great, but does this imply that the future of humanity is merely hinged on building good bots? Is this all we need to stay relevant? Absolutely not. When it comes to preparing for the future, there are certain skills that humans clearly need to develop. 1. Deep EmpathyEmpathy or the ability to understand or feel what another person is experiencing from within the other person’s frame of reference is perhaps the biggest differentiator between man and machines. This is also a trait that is extremely difficult to replicate. Developing a deeper understanding and respect for other’s values, cultures and emotions is the only way in which we can stay relevant by continuing to harness technology to serve human needs.2. CreativityWhile machines may augment human creativity, but our drive and desire to create will continue to segregate us from machines. If harnessed and channelized well, human creativity can ensure we can stay economically relevant. 3. Education With widespread automation of routine tasks, we need to give our children an education that equips them during the Fourth Industrial revolution and essentially implies a complete overhaul of existing systems. Acknowledging that AI & robots are going to “kill” many jobs, Jack Ma, the Chinese entrepreneur and founder of the Alibaba Group, stated at the recently concluded World Economic Forum, 2018 that the knowledge-based approach of “200 years ago”, would “fail our kids” – who would never be able to compete with machines. Children should be taught “soft skills” like independent thinking, values and team-work. (10) Minouche Shafik, Director of the London School of Economics further adds that research skills – the ability to find information, synthesize it, make something of it will become equally important as soft skills.4. Judgement and accountabilityEven if society becomes willing to accept a computer-generated diagnosis or decision, it will always expect a human to be accountable for the consequences. We, therefore, need to train our children in making sound judgements even when faced with complex challenges or situations. To illustrate the above principles, consider the case of Koko bot. This is an anonymous community of supportive people, in more than 150 countries, who talk to one another through chat platforms when people are experiencing depression, stress or just need some advice. The chats must be monitored by humans, to help recognize mental health emergencies. Koko bot still uses human moderators but 90% is moderated by a smart AI that responds with empathetic statements, and facts, to specific keywords. According to a study done by MIT and Northwestern University, 166 people who had symptoms of depression found the Koko Bot to be helpful. (11) Artificial Intelligence can have a profound influence on the way we lead our lives and by being mindful of this before we automate new tasks, we can carefully amplify its benefits. The United Nations estimates that by 2050, the world’s population will reach 9.1 billion and most of this growth will come from developing countries. However, for us to be able to feed this population, food production (net of food used for biofuels) must increase by 70%. Annual cereal production must rise to about 3 billion tonnes from 2.1 billion today and annual meat production will need to rise by over 200 million tonnes to reach 470 million tonnes. (12) Is there a way to be more efficient in food production? As it turns out, AI technology such as FarmView is helping researchers discover the correct genetic makeup of seeds with the highest yield, the most nutrition and genetically engineer the most disease-resistant strains of staple crops. For instance, there are 40,000 varieties of sorghum, a cereal crop in developing countries such as Ethiopia and India. AI can be used to experiment with these varieties to develop the perfect crop. (13) There are other ways in which AI can assist farmers to maximize the current output, help in the precise application of herbicides, give early warning signs of crises or plant diseases and to analyze soil conditions. This is a brilliant example of how AI can help us fight the most pressing challenges of mankind by complementing and even enhancing the ability of humans. In conclusion, so long as machines lack traits like deep empathy, creativity and self-less love, they can never replace human beings – who will continue to find occupation in innovating for a better world, one with equitable access and sustainability. But symbiotic intelligence and synergy is only possible if conscious efforts are made on both ends. Inventors need to adhere to certain basic tenets of AI design and humans need to embrace our forgotten values. At the risk of being prescriptive, failure is not a luxury our generation can afford – it would only mean a doomsday for our children.On the bright side, isn’t our collective quest for a better world at the core of why we’re spearing the AI revolution? How then can Baxter not be a friend?