From a state-run secondary school in Srinagar to IBM’s Watson Lab and now the Director of Engineering (AI) at LinkedIn, Dr Faisal Farooq has watched artificial intelligence crawl out of the laboratories and go mainstream. He has personally contributed to the process. In a long interview with Masood Hussain, the top AI scientist offers an idea about how the world is slated to change completely in the near future and why the quantum shift has a brighter side that humanity must look at
KASHMIR LIFE (KL): Coming this far, what are the further goalposts of Artificial Intelligence (AI)?
DR FAISAL FAROOQ (DFF): Before answering this question, I would like to say that AI is not a new thing. It has been in process since the 1960s. It started around the 1970s, when John McCarthy, one of the scientists at the IBM research lab, coined the term ‘Artificial Intelligence’. Before I joined LinkedIn, I also used to work for International Business Machines Corporation and dealt with academia in Qatar. Also, I did my PhD in Artificial Intelligence in 2008.
So, I have worked in the same laboratory where the pioneers in the field invented the AI technique. Almost forty years ago, they had devised a technique by which the address of all the letters that were posted to America was not being read by any human being. All those addresses were read separately by an AI algorithm and then slotted by putting a bar code on it. This would decide where the letters are to be delivered. That was the beginning of AI.
Now, the question is, how is AI getting so prominent these days, if it has been around since then? This is because AI has gone mainstream. Earlier, scientists used to work on AI. Now so much work is being done on AI and there are many reasons for that.
The first and major reason is that there is a lot of data on the Internet that has been collected for so many years. Now, the machines, and computers that we have can consume all that data. But, at the same time, the power of computers should be high enough to consume all that data.
When I was graduating, the computer that I had or the big computers of the entire university that we had were used for training chat-bots. As of now, I have a more powerful machine in my phone, compared to those, which also signifies that computing power has also increased.
Another reason is that AI algorithms have become much more powerful with detailed research. And by utilizing it to an extent now, we see that it has started understanding natural language as well because of the data that we provide to it. For these reasons, AI has turned mainstream these days.
One of the believers in the AI of this era, Andrew Ng, who has been one of the leading professors at Stanford University for a long time, says that there is no need, for humans to understand or work on something that consumes more than ten seconds, and that work can be done by AI. All the repetitive work can be done by machines and humans will have to do nothing. It will have its benefits, but consequences as well.
The benefit includes that the work will be efficient. Where four people used to work on something in five days, now it will be done in half an hour by a machine. Things and objects will also get cheaper. The negative outcomes include that some of the jobs will go. Unemployment will increase, but at the same time, different types of jobs will get visible.
In my understanding, at least in some of the developed nations like the US or Europe, long-term logistics or long-distance logistics, such as delivering goods from one place to another is a big industry in it and involves many workers. In my view, this industry will not employ people further and will be completely automated with AI. There will be no need for humans in it, except for the last-mile delivery.
Education has come to a watershed moment and the reason behind that is that when we studied in college, we gained knowledge from books and that is also quite valuable. But, now all this will be over. Nowadays, we have chatbots and by writing a sentence or word on it, we can get all the required information or knowledge about anything. But, this knowledge will have no value. Only the skills we have will have worth.
Another aspect that will be changing in our education system is that there is so much information. How can students understand what is right and what is wrong? If we pose a question to Google or a chatbot, it will give an answer with confidence, even if that would be wrong. There will be a lot of things that will change the whole system. Like the industrial revolution changed the world, this will be another watershed moment that will change the world.
KL: By the way, what were the learning curves and experiences throughout your journey?
DFF: Growing up in the 1990s, the conditions were much worse in Kashmir as compared to now. The schools used to be closed, and I used to study at home or go to tuition. I did my schooling at New Era Public School. After that, I qualified for high school at Jawahar Nagar Government Higher Secondary. Then, I wrote the engineering entrance examination and fortunately got the number one rank in Kashmir. I wanted to opt for computer science, but unfortunately, it was not established at NIT Srinagar at that time. So, due to my rank and desire, I was sent to another NIT. Eventually, I graduated from NIT Bhopal in 2001.
I had a belief that after graduating I will be an expert in this field. When I graduated, it felt as if I had learnt nothing. Then, I started applying to US Universities for further studies.
However, I spent a year in India and worked at a software institute called Infosys Technologies. Whatever I earned there was spent on applications and fees for the university outside. Ultimately, I got admission for my PhD at the State University of Buffalo, New York through a fully-funded scholarship.
During those days only, I came in contact with IBM AI Watson Lab, where all essential innovations have taken place. From the making of hard disks, RAM, and Copper conductors to Language Models, all these progressive inventions were made in the same lab. Joining that Lab was a turning point for me and the time when I felt that now I am going to learn something.
When I started research, I published a lot of good papers and met good mentors. After my PhD, I got a job at Siemens Healthcare. Later, with my developing research, I was nominated as the general chairperson of the Knowledge Discovery and Data Mining (KDD) Conference.
In 2019, I went to Qatar and started the Department of Health AI at the Flagship University of Hamad bin Khalifa and then came back in 2021. Later, I joined LinkedIn and became the director of Engineering AI. During this whole process, I learned that one should never get satisfied in the learning process. Being, dissatisfied with learning helps in acquiring more knowledge.
KL: What were the major takeaways from your PhD?
DFF: My PhD was about and around AI. When I was doing my research in the IBM lab, the Google search engine was becoming famous, and I saw that if you search on Google, you can find all text work on web pages, but you cannot find or search for handwritten documents.
The entire work that I dealt with during my research in the lab was handwritten. So, I wanted to connect these two things and see if it can be extended to address interpretation, which is a small part of handwriting, and also make Google search information from handwritten documents that include a vast amount of knowledge. So, my mission was to connect these things.
Also, I applied many techniques and language models to this research, that if there are some errors in the research information, how the computer will correct it. So, I worked on image recognition, language modelling, and search in retrieval, to make the AI use proper. A lot of papers from my PhD research also got published and finally, in 2008, I completed my PhD. After that, my work and those algorithms were applicable for a very long time. A lot of people applied and used those joint learning algorithms.
By now, however, a new type of mission learning has developed, called Deep Learning. Deep learning has changed this field in a wider range. This field has evolved a lot with machine learning because they are more powerful now, compared to back then, when we did not have that much data and computation. Somewhere down the line, however, those algorithms are still bases for the new ones.
KL: What are your views regarding the cycle of knowledge acquisition and technological progression that are reframing the world today?
DFF: The industrial revolution dictated new world order created economic differences between urban and rural, widened income and wealth differences between high and low-income households, and different views on public policy regarding the economy. That is called the great economic divide. With the advent of the internet, education and knowledge is shared throughout the borders. But the divide exists and it is called the digital divide. This divide will get established much faster than the economic divide.
My perspective is that the sooner the digital divide will be established, the faster it will be concluded. And the reason behind that is wealth creation and management can be controlled but information creation and management are very hard and cannot be controlled.
Nowadays, we have a certainty that if a particular bit of information cannot be searched on the internet, it does not exist. And AI algorithms have played a very negative role in it. For example, Facebook has specific artificial intelligence algorithms. And it makes money by showing ads. As long as one stays on Facebook and keeps scrolling and wasting time on Facebook, more money will be scored by it. In a way, it forces you to spend time on it by catching and training things by AI that a person likes, it will show him that particular information again and again. It pulls you into a corner from which you cannot get out because whether you are receiving the right information or the wrong information, you have entered your echo chamber in which you are seeing the world in a way you want to see it. So, if you have a political view, religious view, or cultural view, it will pull you into a shell that may keep you away from other shades of opinion and views. Because of these things, people’s mental well-being has been impacted and because of this, I think more people will break this information control.
This business model will not last long because the incentives are just to earn money and there will be a time when this model will break and people will not like it further. If we talk about Pinterest, LinkedIn, or other social media, they are moving towards a positive side. They are planning how to give information to people that they want, like, and add positive value to their lives.
KL: Do you think that the digital era is pushing us toward a new evolution?
DFF: If we see, this idea has two aspects. One aspect is that evolution has been taking place in humans over time. It is not a matter of today. It goes supplementary with time and environment and you go with the flow of evolving without even knowing it. It is fundamentally slow in nature.
In my young days, we were not allowed to use a calculator in our examinations, because we were trained for calculations using our memory. In today’s world, if we see, do we need to do those calculations now? Maybe at that time, it was necessary to do those calculations with our brains because we lacked the resources, but now we have enough innovations to figure those things out.
It is also a type of evolution where some corners of our brains will not have to be used and some corners will be of more use. Before technological evolution, man used to work a lot. After the advent of cars, people stopped walking. This also changed the physiology of humans. Obesity and other diseases or disorders increased. This is also an evolution.
Establishing AI will not only impact physical evolution but, mental evolution in humans as well. Another aspect of it is that at the end of the day, human beings have a soul and this soul has basic tenants that cannot be ignored. No matter how much AI progresses, you have to drink water and eat food on your own. These basic tenants form the core of our wellness and we need to take care, of how not to get them impacted.
With upcoming new generations, we need to get them aware of our basic tenants and how to resist social degradation because, eventually, humans are social creatures. This social degradation has happened because we can write anything, fight or abuse any person on Facebook or social media but if that same person is made stand in front of you, the attitude will be different, or you will not even prefer to talk to him/her the way you were talking to them on social media. So, what we need to emphasize is our mental fabrication, otherwise, our character will not remain human.
KL: What is your opinion about the new generations being dependent on AI?
DFF: If we relook at the times when I was in school, I didn’t have a computer and had to gain all the basic knowledge from books, had no electricity, and had to use pen and paper. Today, students have become dependent on AI products, like mobile phones, laptops, tablets, and computers. But this is not a new thing. First, we were dependent on books, and then we became dependent on the Internet. We were dependent on ships and now we are dependent on Airplanes. These things have progressed with evolution and will further happen.
These evolutions have certain negative effects but at the same time, they do benefit us. Also, mentioning the impact it will have on education, as I said before, our education system has not changed for many centuries and this is what will change it for the benefit of further generations. We still make a child write a five hundred words essay and give him marks on that basis. The question is, what do you want to teach him by writing? Will it be useful for his job in the future? So, this learning construction and concept will be changed with AI. It will only be the skill that will matter. Ultimately, our education system will have to change because of this. Then the definition of what we call incentivisation or meritocracy will be changed. Because what will eventually matter is how quickly and efficiently a person will work on something.
KL: What is the future of social media, as far as AI is concerned, and what is currently happening at LinkedIn?
DFF: All organizations including LinkedIn have a philosophy that the social media that we are creating should be a positive experience and a safe platform for people. But certain platforms have become dangerous, where people mostly come out with negative sentiments. Social media is slowly evolving into artificial interactions and fake representations.
In my opinion, for social media to continue to be a success, it has to be closer to real social interactions than the fake ones that are being created. Content should not be viral, and a safe space should be created. Also, the hype of being on different types of social media apps has no value. To some extent, it connects us with others and our loved ones too but it also creates a sense of noise and vulnerability, which has a negative impact on us. So, to stabilize social media, we need to have more natural social interactions than artificial ones.
(Umaima Reshi processed the interview)
