Artificial Intelligence- A Friend or Foe for Humans
Technological evolution is always a neutral phenomenon. It's the human which makes the phenomena positive, negative, friend or foe; when Newton discovered third law of motion it was not meant for developing weapons. Same goes for AI - people can use it for developing medical diagnosis equipment or killing machines; the only thing that matters is who is using it and for what. For example, in the medical field the AI algorithms could be used to learn the best chemical constituents that can be used to cure diseases. On the other hand, AI can also be used to develop UAV's to kill people (good/bad) by learning various war skills.
After the machine revolution of 19thcentury, I consider AI as the next big phenomenon in human civilization which will affect the global socio-economic condition. In the 19thcentury we tried to develop machines that can mimic the physical strengths of humans or animals whereas in the 21stcentury we are trying to mimic/develop decision-making capability of humans/animals. Now, some of the mind-boggling questions which can be asked around AI are: Will machines outperform humans in making decisions? Will AI kill jobs or will there be new verticals in the job market?
Will machine intelligence replace human intelligence?
Before answering this, we have to understand the fundamental difference between the two basic activities which the human brain performs in daily life: prediction (what will happen or what it is) and decision (what should we do). So far in the job market, AI will replace those jobs easily which are more of predictive in nature (off-course, with data) for example illness diagnosis, financial anomaly detection, intruder detection from video sequences etc. But for mimicking the decision-making capabilities of the human brain, there still a need for much advancement in the existing AI algorithms.
Having said that, AI's decision making capability in a constrained environment (in games, sports) will soon start outperforming the humans.Normally we humans perform decision making in unconstrained environments i.e. decisions that could be influenced by an infinite number of parameters and therefore it's bit tough for an AI (at least for the current set of published algorithms) to replace humans in unconstrained decision-making areas. Apart from that we humans don't learn just from cold hard facts (like the current set of AI algorithms do) instead we change our opinions over a time and so does the decision making.
In the current form, AI will remain more assistive in nature.It will assist us in refining the processes and adding efficiencies and even though it will kill jobs, it will create new job opportunities as well.
Future Jobs in AI domain
The first and most important requirement of an AI is the data.Data is as important for AI as clay is important for bricks.In the near future, data collection is going to become more difficult because users will become more possessive about sharing their data. To handle such situation the AI domain will need data creators (which of course will not require any high skills) who will generate synthetic data to train the AI. Amazon's Mechanical Turk is one such platform used for creating data by paying people. Even if somehow companies manage to get users' data then also there will remain certain domains which do not have enough data to train the AI. For example, the Cambridge-based startup Audio Analytic employs people to smash glasses and collects the sound patterns which they further use to train an AI security system that could distinguish between the sound of window glass broken and glass broken in the kitchen.
Apart from this when engineers and researchers were developing computers in the 50s, no one knew that there will be "viruses" as well in future which will affect the normal working of computers.In the same way, once the AI advancement reaches a certain stage, further development will be split into two verticals, one that makes advancements in the AI field and one that tries to develop technology to dodge these advancements.
Centuries ago, almost 90 percent of the human population was involved in agriculture whereas today only 2 percent of world population is involved in the same even though we are producing and feeding a much larger populous. We developed instruments, then machines, then computers, then software and now AI and in every such development we killed some jobs but created newer ones.
Humans are programmed to evolve with the evolving environment and every technological innovation challenges us to push our boundaries and creates new opportunities and friends as well.This time it shall be AI.