Artificial Intelligence

Confession - This is primarily meant for those who think Artificial Intelligence is something like artificial insemination or in that zone. Done in the brain area for enhancing chances of harvesting badly needed and inherently scarce intelligence.  Those who have more than a pea-sized understanding of AI are requested to pass this by.  Or if they read this anyway, they are requested not to tell me `ah, this is too basic'.  It is, they are forewarned.  It is a common man's look at AI, not for those who are looking for research material on AI.

Laymen's idea of artificial intelligence (AI) today varies from vaguely amorphous to seriously flawed to reasonably good, depending on how much effort they have made in this pursuit of comprehending the phenomenon. One friend who is ludicrously disinterested in and almost offended by all such things asked me innocuously how exactly would doctors perform this procedure!  Another one who is viciously antagonistic towards politicians of all hues wondered whether an `one kilo package of AI', procured at a discounted bulk price, can be attached to all budding and senior politicians so that they work marginally better, for the benefit of people! As an afterthought, he recommended a higher dose - 20 kilos - for the likes of a few specimens who have risen be presidents or opposition leaders in some countries.  Now I don't know the answer to that because AI's ability to enhance the native intelligence of people, especially politicians, might be more than doubtful.  A few are somewhat better informed, thanks to inputs from people with higher skills and knowledge.  I realize that AI is making things better for the common man in various areas like organizing data, in-depth data analysis, quicker and more informative responses to queries.  As it is making serious inroads into the functioning of conventional technology companies/platforms as they help in manufacturing, defence and other advanced sectors.

But experts don't seem to have fully resolved the more serious question as to how much AI will do to deprive people of their software/programming jobs.  There have been serious differences among technology pundits about this issue.  Some are vociferous in saying a big chunk of job losses will result.  Others believe while people will be thrown out of current jobs, many might be fitted into AI related work that will be booming.  While AI is already being used in many fields successfully for productivity and creative benefits, doubts persist elsewhere about AI being always accurate and correct. So, in critical domains like medical science, AI may have to go much farther for convincing results. If you wonder why these questions have not been popped to AI for clarity and resolution, I am with you!!  Another spin ball that has dogged the experts is how much investment into AI related build-up is too much!! Some companies were overly gung-ho in the last few years in piling capital investments into the domain of AI. Of late they seem to have pulled back a bit, after realizing that pit is bottomless and sustained returns may not as assured as they thought. 

In this swing from zero to aggression to caution, these companies have lost market-caps and investors have lost their pants in the stock market. Very dependable, marquee technology companies have taken big hits to their bottom-lines, profitability and also share prices in the past few years. Principally because there was no clarity about the impact of AI on traditional software/IT companies' business prospects and their robust profits. As was most likely to happen, these companies are making phenomenal adjustments to their own business models and plans, to accommodate and gain from AI as an integral tool in the way they build technology platforms for themselves and their clients.  Very soon such technology behemoths may metamorphose into IT and AI companies - such a synthesis will happen eventually.

Remember the early days of automated spreadsheets?  They integrated many formulae and made calculations easy.  Calculators and mental computations became passe overnight. That was for just one minuscule task, but that benefited a section of humanity immensely by saving millions of man-days involved in computing, accounting and tabulating.  Now, multiply that a billion-fold or more and apply to any subject under the sun.  This is what AI is aiming to be possibly doing and the results will be staggering.  Before the advent of AI, we were popping questions to Google or Bing and getting simple one sentence answers (or more) on any subject, which had related information compiled and stored in the cloud.  With AI, the quantum of information stored as well as the capacity to analyze the same in a jiffy have increased exponentially and the outcome is that very complicated inquiries even from domain specialists/experts are getting answered effortlessly in a blink. So enormous amount of time is being saved and obviously productivity is increased multi-fold. This comes from the fact that the quantum of data available with AI is huge and the logic built into it for analysis is very advanced. 

One thing must be clear by now.  The quality of the responses generated by AI will entirely depend on the quality of the information already stored, ongoing updates and the correctness of the analytical tools used.  If the information assimilated is compromised or the analysis is flawed, the answer is going to be goofy, unreliable and harmful too.  Recently a report came out somewhere that people should not blindly trust the answers they get from AI in the area of medicine and go for treatments. Because there seem to have been instances of the analysis being erroneous, thereby rendering the entire response questionable.  What this means, at this stage, is one cannot use AI to save one's life because what it puts out as a solution may be far from a life-saver.  Such a response will require manual intervention, for a review by an expert. Meaning, one has to still go to that painful doctor for diagnosis and treatment; one cannot self-medicate, something Google has always told us.   

Where will all those poor souls who are likely to lose their jobs go?? Especially if software coding itself can be done using AI by anyone, without knowing any coding language.  Redemption is around the corner, may be. All data compiled by AI from multiple sources should be seriously checked for accuracy for obvious reasons. By whom? You should not be surprised, primarily by us humans as of now.  So, compilation of huge amounts of initial data gobbled up by the AI platforms, updating that constantly, categorization, labelling, checking for accuracy, editing, corrections and finally implementing the last step to get AI to be ready to answer fresh queries have all to be accomplished by people. Even though the AI platform itself can pull in new data  and to some extent do the editing/correcting/updating, I understand.  But, leaving AI as the final arbiter in this process will not inspire too much confidence, at least for now, in people.  This means a lot of manpower will be shifted from technology jobs to AI work.  That will have to continue, until AI platforms are fully self-sufficient and can manage all the above tasks by themselves without the intervention of anyone.  Technology personnel will have to be re-trained for different things and for higher level intervention.  Something like, a cook making shapeless onion pakodas now will have to be readied for gourmet Italian food.  Of course, the world has to develop faith in the whole AI process, its integrity and the accuracy of the output.  This will take time.

AI surely is being asked more heavy duty questions - more than mundane questions like `what is the best route for Bangalore to Chennai', which we reserved for Google and Bing.  Certainly AI should be benefiting people at large and youngsters especially.  But, it is not clear how much it would help students to skip classes completely and rely on AI. A certain segment of students might do that because AI can become a substitute lecturer, as we have seen in a TV commercial.  Should it matter, is a question for the academics to sort out.  If learning is the ultimate goal and students do that well from AI rather than the mostly boring classrooms today (at a fraction of the cost??), will that be an issue? They might miss the essential fun part of student life and social and real-world skills.  And then there are individuals like the ones in an another TV commercial who asks AI `I have spent half my salary in shopping and still have to go for 25 days for the next pay cheque.  How should I manage'?  As if you need AI to tell you that and this is the real danger.  So long as this is not the prevalent scenario among the majority, the so-called Gen Z, which is hoping to revolutionize everything in India should be fine.

When I asked my dear wife what she thought of AI, as a true-blue teacher, she said it should be used extensively to supplement teaching methods, especially to offset the abysmal quality of teachers in general.  But she also wondered, based on past empirical evidence, if despite the prevalence of AI and its monumental analytical prowess, people would progressively become naturally more stupid by the day because of over-reliance!!  Possible.