Sci-fi fiction and movies have come up with scenarios in which humans are dominated or controlled by androids – computers behaving as humans. Some scientists feel that these fictional scenarios are highly probable even within a few decades – say, fifty years from now. Based on the way technology is growing, let us try imagining a realistic scenario, keeping in mind that fifty years is a long time for predicting as dynamic a field as computer science and also that major disruptions of a positive or negative type can result in a very different scenario.
For the sake of clarity and uniformity, let us call the human-like robots as androids. Even if androids do not have the psychological and spiritual elements of humans, they should be able to exhibit general intelligence like that of humans. They should be able to perceive the environment, solve problems, reason, and understand and talk in human languages.
How dependent have we become on technology
Starting with their earliest avatars, computers have been helping us do things beyond our limitations.
In the initial days computers were available in tech labs only; for ordinary people, electronic gadgets like calculators and electronic typewriters and early social messaging systems such as bulletin boards too made the life much easier with calculations, typing, sending messages and such activities.
The early personal computers had very little processing power and storage, but they looked out of the world, with their capacity of text and data processing, image viewing, etc. Then came pagers, mobile phones, internet…
Within a few decades, ICT (information and communication technologies) have grown thousands, sometimes even billions, of times in efficiency. Over time, these technologies became affordable for the common people, and now they have become an essential part of life for a big majority of people.
Machines (starting with mechanical, then electrical and then electronic and computerised) started making production and transportation more efficient since eighteenth century. These started supplementing and then replacing workforce, especially as the industrial age set in.
Talking of computer power vis-à-vis human brain, we have developed computers and software that can defeat chess grandmasters and top players of Go (a highly complex Chinese game). They can solve the Rubik’s cube in a second.
Earlier the machines were dumb, but now they are becoming more and more smart and autonomous. You can keep your aeroplane in autopilot mode and take a quick nap. Robots are being used for trash collection and security. Driverless cars are the latest example of machines that can make decisions on the spot based on traffic, approaching human beings, road condition, road alignments, traffic signals and so on. In a very narrow sense, these machines act like humans.
Much of the decision-making as mentioned above is possible when the programmer can tell the machine what to do when a particular situation arises. When the program and associated hardware can take care of all possible situations, the machine seems to perform as good as or better than humans.
Machine learning plays an important role in decision-making by machines in a dynamic situation – situation which the system has earlier handled based on the program, but some parameters have kept changing. In this type of system, beyond a base logic, the computer is able to learn from past experience.
A commonplace usage of machine learning is in understanding the intent of queries by search engines. As more people search for a particular thing using different search expressions, search engines get better in knowing the intent of the searcher. For example, when somebody searches for apple on Google or Bing, the search engine utilizes its experience to sense whether the query is about apple the fruit or an Apple gadget. Machine learning is behind Facebook and YouTube serving you videos based on your search, browsing, and social and e-commerce behaviour. Their algorithms also use machine learning for spotting fake news. Your email provider filters spam efficiently based on machine learning.
Scientists have borrowed the architecture of human brain to create artificial neural networks. This helps them create systems that can take diverse inputs and then classify them for better machine learning, etc.
Recall, how we forgot the art of remembering phone numbers, etc. once these were easily available on mobile devices. Then the same happened with locating addresses. When we have to visit a new friend’s place, we now seldom ask her for the route to her address; we put the navigation on while driving and reach there by the shortest route.
Digital assistants (e.g. virtual assistants: Siri, Alexa, Google Assistant; assisting devices: Google Home, Amazon Echo) are taking our dependence on technology to a new level. Something comes to your mind while brushing your teeth, you speak that out to the assistant, and you get the information instantly. The assistant plays songs for you, connects you to people, deals with calls, reminds you of important assignments, and controls other devices.
Devices such as assistants (including virtual assistants on computers and mobile devices) understand your language and reply to you orally. In the coming years, they are likely to sense your thoughts by gestures and even based on signals generated by your wearable devices.
Internet of Things (IoT) is not much common in India yet but is becoming mainstream in developed countries. In IoT, the smart electronic gadgets are connected to a controlling system and can be given direction remotely or depending on a particular condition (e.g. the smart AC at home can turn on when the owner has started driving home from his office).
Robots are not new, but they are becoming sophisticated by the day. They are being extensively used in industries and trade, so much so that in some areas they have completely replaced human labour. Robotic gadgets are already in home use (e.g. robomops, robotic toys).
Machines are becoming smart but we are the masters
Right now, Artificial Intelligence or AI is being used in many applications, only some of which are directly used by the common people; much of it is used in technology and industry. Even then, we are getting overwhelmed by technology. If you use the gadgetry and software available in the market, your home is fully smart, and most devices can be remotely controlled. They automatically sense the right temperature, humidity, sound and light levels for their correct operation. You have robotic servers to serve you food and drinks. Based on your consumption, the fridge and kitchen cabinet order groceries, etc and make payments. Your wearable device is there to remind you of your breakfast time, office time, meeting time, giving a birthday wish to someone, and so on. It also constantly measures your health parameters, orders medicine or books an appointment, reminds you when to take medicine, and cautions you when you binge on some food or drink.
An automated machine does vacuum cleaning, mopping and dusting jobs at your home and office (but you are not satisfied with its current version and are looking for its replacement whenever better models come that can climb stairs and reach the ceiling). You have to do with a car that has assisted system (since the driver-less car is not yet out) for navigation, hazard sensing, traffic alerts, auto-parking, etc. Whenever you enter any building, the doors open if you have the access ID. On the shop-floor, you still see humans, but they are being replaced by robots. Movement of large tracks of supplies is, of course, fully robotic, much like the machines in your construction project. A lawn mower, robotic security guard and AI-equipped personal drone are on your shopping list, but good ones are not yet available in the market.
You have got some of the machines dressed like smart ladies or gentlemen. Some, if not all, look like lovely humanoids.
If you are not yet in that scene, you will be in a couple of years. There will soon be so much prevalence of AI and smart things that your things at home, in office and in the market/ roads will be in constant interaction to keep you informed about everything you want, facilitate your stay and travel, check bookings/ bill payments, etc for you… In effect, technology will take over your life in a great way – but things will mostly be under your control. The machines will not have taken away the job of thinking and reasoning from you. You might have an assistant or a software that takes notes of your meetings in the conference room and produces a summary of decision points as soon as you reach your desk; but you do not yet have a machine participating in the meeting and hotly arguing his point. You also do not have a machine subordinate who goes on strike when you don’t meet his demands.
Technological inventions in the direction of developing thinking machines has, of course, been slower than what sci-fi fiction has been predicting. But artificial intelligence has great potential. AI is fast, efficient, tireless and usually error-free. AI makes other machines extra-smart. So, the development and commercial production of AI-enabled machines is expected to increase at a fast pace – led by startups and eccentric risk-takers (Elon Musk comes to mind).
Does that mean, machines are going to be our masters soon? That can’t be answered in black and white, because – as I said at the beginning of this discussion – technological developments are too dynamic to be predicted with certainty. Yet, let us think in terms of possible scenarios.
Scenario 1: Androids will become a reality in not too distant a future.
This is the most fascinating and equally scary scenario, drawn straight from sci-fi fiction and movies.
There are some scientists who think that fifty years is good enough time for the technology to reach a level when computers/ robots will almost think like human beings. The progression towards autonomy and automation in computerised machines is a good pointer to that. It should also be kept in mind that once a small breakthrough is achieved, technological enhancements are very fast.
In this doomsday scenario, just one rogue AI machine deciding can carry out enormous mayhem, even eliminate humanity.
Scenario 2: Human-level intelligence is not possible any time soon.
Some scientists believe that intelligent machines that can mimic human intelligence will be very difficult to make. They point out that the concept of artificial intelligence first arose nearly a hundred years back. In 1950, even a test (Turing Test) was developed to see whether a machine could exhibit general intelligence like humans. Progress occurred in many disparate directions, but almost all have reached a dead end. What we have got is artificial intelligence which is just an extension of high-level computing (weak AI) and its smart application, and not human-like intelligence(strong AI).
Though theoretically it should be possible to create androids, it may take many more than five decades.
Scenario 3: Even with a sub-human intelligence, androids will dominate human lives.
This is the scenario that is most likely to unfold in the coming decades. The marriage of artificial intelligence with other emerging technologies will soon create computer systems/ assistants that
- fully understand human languages
- find the intent of human-made queries and instructions to a great extent
- predict outcomes and act/ plan accordingly, based on learning from the past
- analyse fuzzy (with no clear-cut boundaries) situations more correctly
- handle some level of uncertainty
- act autonomously to the extent that they carry out a series of actions and take certain decisions without human intervention
- can respond to environmental factors in making decisions
The use of intelligent machines is likely to expand hugely unless governments reserve some jobs for humans. Such machines are likely to be the norm in all spheres of human lives within less than five decades.
It is also likely that due to aesthetic and other considerations, many of these autonomous machines will take the shape of humanoids. These will have human-like shape, size, body filling and cover (like human flesh and skin). Moving parts and associated mechanics would make them imitate humans in body movements, including expressions. But these humanoids will not be intelligent enough to be called androids.
Omnipresence of humanoids and other semi-intelligence systems will raise serious ethical issues such as unchallenged intrusion into privacy, control over people’s choices by corporates and governments, possible bias against the weaker sections of the society, and its use in warfare beyond the battlefields.
After that, development of full-fledged androids and their domination will depend not only on the technological progress, but also how humanity handles it. There already are scientists and policy-makers cautioning against too much interference of AI in human affairs. It is likely that governments and international organisations come together to control the situation when the tripping point occurs. Humanity has been more or less successful in handling nuclear, chemical and biological armament.
There are some who are concerned that eventually AI will start controlling humans, and insist that research into AI must be guided in a way that it is used only for useful purposes and is not allowed to go out of hand. They include the noted physicist Stephen Hawking and Microsoft founder Bill Gates.
On the other hand, there are some who see the danger, but think that it is too early to lose sleep over it. Let me quote Andrew Ng, a well-known computer scientist: AI existential risk is “like worrying about overpopulation on Mars when we have not even set foot on the planet yet.”
So, as of now, we can rest assured that humans will not allow androids to misbehave, take law in their hands, revolt or subjugate them – not in the next fifty years. If you are beyond 50, you will not see such a situation – let me guarantee you that!
If you want to go deeper into the subject, I recommend these resources available on the web:
- AI Has Arrived, and That Really Worries the World’s Brightest Minds
- Android (robot)
- Artificial general intelligence
- Artificial intelligence
- Potential Risks from Advanced Artificial Intelligence
- Preparing For The Future Of Artificial Intelligence (PDF)
*Manoj Pandey is a former civil servant. He does not like to call himself a rationalist, but insists on scrutiny of apparent myths as well as what are supposed to be immutable scientific facts. He maintains a personal blog, Th_ink
Disclaimer: The views expressed in this article are the personal opinion of the author and do not reflect the views of raagdelhi.com which does not assume any responsibility for the same.