Intelligence is often defined as one’s ability to perceive and interpret one’s environment, acquire knowledge, and apply it in a suitable manner. Humans are considered reasonably intelligent beings, so are animals (did you know pigs have an IQ nearly equal to a 4-year-old human child’s?) and even plants. How about objects? Chairs and tables? Perhaps if chairs backed and tables set themselves without external help every time we tried to have supper. Considering they don’t, though, we’ll rule them out, but what about machines? I urge you to think about the last time your keyboard’s predictive text completed your sentence for you, or Uber or Lyft showed rides near your location before you answered that.
Thanks to extensive dramatization in novels and films, what first comes to our minds when we talk about AI are often the larger-than-life ‘Transformers’ from the movies, stepping on buildings and taking over the planet. Being a smidgeon more realistic reminds us of robots who walk and talk. In reality, artificial intelligence (henceforth AI) is a rather elite technology that usually looks nothing like the above and has gradually become part and parcel of our lives. Simply put, whenever a machine executes the job of the brain, it is an example of AI. AI is built to mimic a human brain by perceiving information, processing it, and taking actions that maximize the efficiency of the output. It can learn and solve problems, as well as rationalize and be goal-oriented. Now our dependence on AI is to the extent that without it, a lot of our daily tasks would be rendered so cumbersome they would linger on the border of impossibility.
What has enabled AI to become so indispensable is its high level of efficiency and objectivity. Being mathematics and computer science based and slowly beginning to incorporate linguistics and even psychology, room for error in AI is minimized manifold, especially when compared to human systems. AI trends have always been in line with the upcoming needs of various industries, and with further development of AI tools, these industries are either already being or can be revolutionized to a large extent. Some of these industries are discussed below.
Since one of the greatest upsides of AI is its ability to look through massive amounts of data in very little time, spotting discontinuities in scans such as MRIs or blood reports has become significantly simpler. The accuracy of results has increased as pathologists have assistance. In fact, it is not just the interpretation of results but also the process of scanning itself, which has become assisted, hence much more simplified, with a good section of scanning technology being automated. This saves time and reduces the margin of error. A trend of early diagnosis of cancer has followed, and in the Beth Israel Deaconess Medical Centre, deadly blood bacteria are being detected at a very fast rate using AI.
When it comes to treatment, another property of AI, which is an interpretation of data and prediction of the future by studying trends using advanced mathematical tools, has made it much simpler for doctors to prescribe drugs or make personalized treatment plans. AI has made it possible to predict the efficacy of a drug or extrapolate patient characteristics and look out for negative side-effects. The chances of treatment not working, though still present, are reduced to a certain degree.
Saving time in diagnosis and treatment has a further impact. Taking less time to diagnose and treat patients means a greater number of patients can be diagnosed and treated per day. Not only does this mean a greater number of people get quicker access to healthcare, but it also points towards greater financial resources in the healthcare system. This increase in funds is reflected in better technology, more number of and superior quality of doctors, more facilities for the patients, etc.
Further ways in which patients benefit from the incorporation of AI in healthcare systems is through the enablement of distant monitoring (which increases vigilance to 24*7 instead of just when the nurse checks in on you), simplicity in taking medical decisions due to the availability of much greater data and predictions as compared to before and a smoother healthcare experience by virtue of streamlining of the system. There has been a reported improvement in patient experience and feedback with paperwork being efficiently managed by AI and the subsequent lack of confusion at the front desk. Intelligent symptom checkers or medical chatbots have even made it possible for people to seek healthcare without stepping out of their homes in some cases. Although remote healthcare has its risks, if used wisely and with discretion, it can be highly convenient.
When it comes to finding suitable candidates for clinical trials, combining AI with the cloud and quantum physics has proven to be of unprecedented utility. Never before has the process of selection and formation of sample sets been as sophisticated as AI has now made it to be. And this sophistication goes beyond forming groups. Making large statistical functions machine-dependent (for example, computerizing tests such as ANOVA, post-ad-hoc, etc.) plotting intelligent graphs and monitory bioactivity and patient characteristics have helped researchers find out the efficacy of their results and materializing thought experiments with modifications. Ebola, multiple sclerosis, and now COVID-19 are only some of the diseases this methodology has helped in. Another beautiful example would be an AI computer known as IBM Watson of the IBM Watson Health company reviewed thousands of research pieces and identified a new gene linked with Amyotrophic Lateral Sclerosis (ALS), also known as Lou Gehrig’s disease in the December of 2016. Since then, the discovery of five more genes has followed, and ALS treatment research has been taken to the next level.
Research has shown that one of the reasons students opt out of higher education or drop out mid-semester is the lack of proper support. Since every student has a different starting point, learning speed, and ability to understand, it becomes difficult for a professor [even with her Teaching Assistants (TAs)] to personalize teaching for each of her subjects. This is when AI comes into the picture.
In 2016, Georgia Tech University of the United States launched distant learning in its Masters of Science in Computer Science program. There was a small experiment included in the program and, spoiler alert: it was successful. The program employed a virtual assistant to help the professor and his eight human TAs, and after some troubleshooting, the assistant started answering student queries without any human intervention. The accuracy of answers was 97%, and the students never found out until the end of the program that one of their teachers had been a robot.
Other forms of technical assistance include automated administrative work such as faster admissions and enrolment procedures and systematized homework correction and test evaluation. AI is developing to the extent of evaluating subjective answers also now, along with objective multiple choice question answers. Differently-abled students or students with language barriers have also benefited due to AI with provisions such as quick and instant subtitles.
Using AI’s advanced analytical abilities has enabled governments to make crucial decisions with greater precision and welfare maximization. Plans about the economy, the environment, and various pressing social issues can be made in real-time. Planning cities, monitoring and understanding patterns of climate change, and reducing poverty using schemes are all processes made more convenient using AI. Further, AI’s quick detection has helped spot hate symbols and hate speech, looking out for terrorist activity and preventing attacks. In the military, it is employed to look for any possible attacks, and upcoming technology also helps prevent missile invasions by countering them mid-air. Cybercrime can be restricted by studying mathematical DNA.
Automatic dimming of phone and laptop screens according to battery power, scheduled screen sleeping, and self-regulation of condensers in air conditioners are all examples of AI minimizing energy usage. While all of these activities might see quite lowkey, when coupled with the sheer number of users worldwide, every reduced quantum of energy seems significant.
In large scale usage, an example would be google recently employing its AI platform ‘Deep Mind’ to regulate cooling fans for the thousands of servers in its massive data center and offices worldwide. Under this technology, cooling fans turn on only on overheating of the device instead of running throughout the running duration of the machines. The energy cut as a result of this has been more than 40%.
Tracking animal movements using AI allows researchers to see where they go and as a result, which habits need to be protected. For instance, this Montana-based study pinpoints the best places to create wildlife corridors – continuous areas of protected land that link zones of biological significance that animals can use to move safely through the wilderness – for wolverines and grizzly bears.
As in the case of healthcare, it was able to analyze massive amounts of data that can transform wildlife conservation.
The next time you are using Netflix, Apple TV or Amazon Prime, I urge you to pay closer attention to the ‘For You’ section, because that is probably the most niche and exciting example of AI interceding user needs so readily available near you. Without knowing about it, AI is continually assessing our patterns of usage and behavior and our likes and dislikes (by what we click on and what we press skip for) and computing the needs that might arise in the future as a result of these things. This is how Netflix gives us suggestions that are likely to keep us hooked to our screens, Spotify curates playlists that we cannot stop listening to, and Zomato knows just the takeaways we won’t be able to resist. So the next time you look up piano buying options on google, do not be surprised if the ads preceding your Youtube videos are suddenly about ‘Simply Piano’*!
*An app that teaches its users how to play the piano (hmm, an example of combining AI with teaching?)
While the utility AI has is massive, there are also some apprehensions in people’s minds about its extensive use, and some of them exist for a good reason. While ‘robots taking over the planet’ is, as an idea, a little too far-fetched, with AI, there are several grey areas that are tricky to clarify, leading to specific ethical issues. The further section is about two such exciting dilemmas.
Although we are still a long way from the actual usage of self-driving cars, they have already become a hot topic of intensive debate and discussions. According to a study by Stanford University, self-driving cars could significantly reduce deaths by traffic-related accidents, making roads much safer. They are also likely to reduce stress levels in people by taking away their cognitive road. However, everything here is not as simple as knowing when to take the right turn. Arguers against self-driving technology have asked the question: if there is a situation where either the pedestrian in front of the car is saved or the rider sitting inside, what will be the car’s decision? When a human is driving the car, she might or might not slam the brakes in the middle of the road breaking traffic rules and jeopardizing her own safety to save the pedestrians in front of her. It would be a moral decision she makes or does not make for herself. However, is it ethical to turn morals into objective lines of code? Is it moral, in and of itself, to universalize morality? And if lives are lost, who will be held accountable; if justice is demanded, who will be tried in court?
The Citizenship of Sophia the Robot:
In 2017, citizenship was granted for the first time in history by Saudi Arabia to a robot- Sophia. Sophia is a life-sized, incredibly smart, and unsettling real looking robot created by David Hanson in 2016. What makes her unique is how her mechanic brain waves are stupefyingly similar to human brain waves- to the extent that she is said to have sentience, a fact that is further visualized by her ability and display of numerous human expressions. What her status should be, though, remains an ongoing debate. Considering she displays emotions to a reasonable extent, is it fair for her to not be treated as an individual entity? On the other hand, if you consider her a legitimate citizen, can she subscribe to the exact same laws as human citizens? If yes, whose vote is it really when she votes during elections- her own or her programmers?
These moral problems- much like most others- do not have objective answers. The debate, however, continues. What is your opinion?
There is no doubt about AI’s ability to revolutionize the way life with technology has been perceived until now. It has in the past, and with further development, can continue in the future to take us to unprecedented heights and probably save and modify thousands of lives for the better. That said, the ethical questions it poses and the controversies it creates cannot be treated by looking the other way. The extent to which AI will entangle with our lives will have to be a combination of our quest for advancement and our moral code as humans existing in this large, fast-paced world.