Technology can be credited to have drastically shaped the world of today. Ever since humans acquired the ability to design things as per their convenience, the ability of humans to constantly innovate has been evolving. From steam engines to fully magnetized levitated bullet trains, from having no ways for distant communications to developing smartphones, we as a human civilization have come a long way in developing advanced technologies. The giant technological leaps are just not mere abstract achievements but also signify our advancement in defining the limits of what human intelligence can achieve. Given the astounding marvels of what human intelligence ranging from inventions, discoveries, creating new art forms,etc,it should delight us to imagine the possibilities in a world where human intelligence gets exponentially multiplied. Artificial Intelligence (AI) can be our gateway to these possibilities.
Artificial Intelligence or AI refers to the technology which is capable of mimicking human intelligence in order to solve tasks such as calculation, critical analysis of data, problem solving, facial and speech recognition and even speech citation. The most fundamental goal of AI is to mimic human intelligence as widely as possible.
Current scenario – if we look at the current scenario of our world, we will realize that AI has already become an inseparable part of our lives.AI due to its enormous benefits finds application in almost all the sectors. From educational institutions using AI for biometric attendance to police departments across the world using AI for tasks such as face recognition. From voice assistants like Alexa and Siri to our android smartphones which assist us in a variety of tasks, we are already witnessing an era where AI and humans are existing coherently.
Recent boom – Social humanoid robots like Sophia which are capable of learning human expressions over time and to have conversations with humans on predefined topics such as weather-signify the boom in technological advancements related to AI. Courses and degrees dealing exclusively with Machine Learning have been introduced in many universities and other similar platforms (Coursera,edx,etc) in recent years, a yet another example portraying the booming field of AI and Machine Learning. One of the best ways to appreciate the boom of AI would be to consider the success of social media giants like Instagram, Facebook,etc.From social media sites to dating apps like Tinder, each app gets its likeability factor from their algorithms. These apps evolve their algorithms by continuously analyzing data patterns of their users.
The analysis is largely done by AI to arrive at suitable conclusions.
However we must remember that none of the devices that we see around in our daily lives like smartphones, voice assistants, etc represent artificial intelligence in its totality. we can understand AI more widely if we divide it into two broad categories.
- Specific AI – This includes usage of AI to perform some specific predefined tasks.Our smartphones,an automated coffee machine, fingerprint sensors are some of the examples of specific AI.
- Integrated AI- Integrated AI on the other hand refers to AI which is built on the foundations of Deep learning.
What is deep learning?
Deep learning involves usage of large amounts of data to pick out and use certain relations and patterns to arrive at conclusions.
This can be better understood using an analogy. During the evening, say 7pm, the sun turns orange and is about to set. A deep learning algorithm will associate 7pm with orange sun without actually knowing the cause behind the orange color of the sun. Deep learning algorithms work on patterns without taking into account the elements of causality.
Is causality always necessary?
Now we might be tempted to conclude that since AI is incomplete and partial in its approach towards causality, it is not a technology which can be relied upon. However there are enough facts and arguments to suggest that causality shouldn’t always be our prime concern as far as AI is concerned. In 2012 Daphne Koller and Andrew Beck from Stanford University conducted a study regarding the ability of machine learning algorithms inspired by deep learning to identify signs of cancer. The algorithms were able to identify 11 signs of cancer as opposed to only 8 signs of cancer known to medical practitioners. This was a big breakthrough in identifying early signs of cancer and at the same time it proved the argument that causality isn’t always necessary. There are many examples from our day to day lives where we overlook causality-we take Aspirin without knowing how it works in our bloodstreams. A major drawback with our obsession with causality is that causality requires a lot of experimentation, trial and error and even after that it is uncertain that true nature of causality can be known. Perfect example can be the evolution of laws of motion. The causality of laws of motion evolved drastically over time. From Aristotle to Newton, the causality(actual reason for something) regarding laws of motion changed drastically over centuries, signifying the large amount of time and labor required to fathom the actual causality. AI technologies which make use of Deep learning derive conclusions which might often go unnoticed from humans. Deep learning make inferences from a manifold of patterns thus ensuring minimum deviation between conclusions and reality.
How AI will affect employment?
A major concern related to the advancement in AI is that of unemployment. With more and more industries getting automated, there are opinions that it will skyrocket unemployment. While the fears of unemployment may be true in the short run they don’t represent the absolute future. The historical Industrial Revolution changed the face of manufacturing processes forever.
Manufacturing processes changed drastically and became more efficient. Goods that had once been painstakingly crafted by hand, started to be produced in mass quantities by machines.
Various industries like the Iron industry benefited largely due to the increased productivity and efficiency which was offered by the machines. Production process became much faster and required far less human labor. Industrial revolution changed the mode of employment from manual labour to working with machines. The industrial revolution can be considered similar to the possible era of AI in the sense that it provided an opportunity for change in the mode of employment. Will we look back in history and argue that we should have resisted industrial revolution because it changed the mode of employment? IAI technology might cause unemployment in the short run in some sectors, as industries become more and more automated, however this situation is not going to persist in the long run. Unemployment can linger only if the population is unskilled in AI and machine learning; the government has started taking due care of that- by introducing initiatives like “Skill India” in order to make people ready for future technologies such as AI and machine learning. In the near future humans working with advanced automated machines will become the new normal. AI will boost the performance of companies in understanding consumer behavioral task which requires enormous calculations. Thus we can argue that AI will boost our economies rather than being a threat to them.
Major challenges in acceptance of AI-
- Racially biased nature of AI-one of the most surprising drawbacks of AI is that it can be biased.
AI afterall is based on a set of programmes and algorithms which are designed and created by humans. There is enough evidence to suggest that human biases can express themselves in codes and algorithms. There are numerous instances where the AI systems have failed to recognise black people as opposed to white people.Around 2015,e-commerce giant Amazon
recognised that its AI algorithm which reviewed job applications was biased in its approach towards women. The algorithm downgraded the applications which contained the word “women’s“. However experts have argued that this biased nature of AI can’t solely be the grounds for rejection of AI. The biased nature of AI is nothing more than certain fallacies and flaws in the algorithms. These faults are computational and can be corrected,so rather than blaming the technology to be biased we should blame the society for having biases.
- Data privacy- this is yet another concern which has been raised time and again.
As AI becomes more efficient and advanced,it is becoming easier to carry out cyber attacks which are capable of hacking entire systems and hence steal vital information and data. In this age of digitisation people are leaving various digital traces without even knowing. These traces are being continuously maintained and analysed. There have been allegations that voice assistants like Alexa and Siri continuously record voice inputs without users even knowing. Whatever we say after the ‘wake word’ is recorded and sent to company servers. The problem is not with voice data being recorded but with voice data being stored by companies for analysis and improvement of voice assistants;a classic example of user privacy being compromised. While these issues are definitely serious concerns, it can be argued that these instances portray the misuse of technology. Antivirus softwares can protect systems from being hacked and can prevent data leakage. Thus it can be argued that AI as a technology is not intrinsically good or bad. Use or misuse of technology depends on us and hence for issues such as data mining and cyber attacks- development of strong AI protection systems can offer appropriate solutions. Moreover strict cyber laws can be enforced by governments in order to prevent cyber attacks from happening.
Will AI supersede humans in terms of intelligence?
A majority of people have fears that AI will supersede humans in terms of intelligence and hence will become a threat to the very existence of humanity. While AI has definitely superseded humans in some aspects of intelligence like pattern identification, calculation ability,etc we must remember that these abilities do not exclusively form the human intelligence. A large part of what makes us human comes from our elements of self awareness and consciousness. “Theory of Mind” suggests that if machines/AI have to become like humans they need not only to become self aware but also become aware about the feelings and emotions which other creatures in this world can exhibit. AI will have to acquire the ability of adjusting its behavior according to the feelings of other creatures during the interactions, the possibility of which happening is almost null. Another concept which points to the large gap between humans and AI in terms of integrated intelligence is the Moravec’s paradox which says that it’s very easy to equip machines with abilities which are very complex on human part to do such as pattern identification, critical analysis of numerical data, playing games like chess,etc but it’s very difficult or impossible to equip machines with abilities which humans find very easy such as precise movements, proper perception,etc.
The road ahead-
The road ahead is definitely good for AI. With the help of initiatives such as the “Skill India” the government has already taken steps to make future generations ready for the change. We need to acknowledge the fact that much of the resistance which AI is facing is largely based on assumptions which may be far away from the reality. We need to be aware of the fact none of the major problems we are facing today like poverty, inequality, border tensions,etc are a consequence of AI. As mentioned earlier, problems like racially biased behavior of AI and weaponizing of AI aren’t intrinsic problems of AI but rather signify failure on our part to be able to create a society which is based on equality and where wars don’t predominate.
AI as a technology has the potential of propelling us into a new age of automation and digitization. Complete embracement of AI can be ensured if certain rules and regulations are made to keep AI within ethical boundaries of humanity, the responsibility of which lies on global leaders and experts.
With great power comes great responsibility and thus it’s the responsibility of humans to ensure moral usage of AI.
-Vishal Kumar Singh
Support the author and follow on – Vishal Kumar
Participate in our art contest and win prizes up-to Rs 12,000 Click here