We have gone from a processing-limited society to a data-limited society. Theoretically, if you had infinite data, you could get the machines to learn everything. Dr Palanivel Thiaga Rajan, Hon’ble Minister for Information Technology & Digital Services, Govt of Tamilnadu, was the Chief Guest. The book has been authored by CA Pattabhi Ram V, public speaker, teacher, and partner, Yoganandh & Ram.
Dr PTR Palanivel Thiagarajan
The impact of AI today is remarkable. AI is not a new technology. Large language models have been around for a long time. This is probably the third or fourth coming of AI. But what is different now, is its sheer capacity to collate and aggregate data; an almost infinite capacity to store it and the on-demand computing power, scalable minute by minute. We can run almost any kind of analysis we need. This was not possible in the olden days as we couldn’t afford to buy the computing power. The notion of cloud computing has really enabled those who need a high amount of computing power, for a short period of time, to really be able to do things in a cost-effective way. This was simply out of reach, because of the capital investment needed to get that kind of permanent capacity. AI is already starting to change our lives. Not only do students use such aids, but teachers have a tool to check if they have used these kinds of aids, when students submit their assignments. The applications, of course, are endless. I often emphasis its use, particularly, in medicine and law. I think these two are areas of immediate application. In medicine, I’ve seen a few companies that take somebody’s coughing into a phone and diagnose if they are affected by TB or some other respiratory diseases. In the law, it is all about the database. There are engines that can not only aggregate previous decisions and show precedents but take it to the next level.
Precedents influence judgment
I served on the GST Council. I was on the subcommittee on horse racing, gaming and casinos. We were discussing how we should tax these. In every meeting, some three or more precedents would come out of some interested party, that were basically conflicting all the precedents we had heard before, through high court or supreme court judgments. This notion that precedent sets law and that everybody reads every case first, before they write the next judgement has been around. We found it very difficult to do our jobs, because we had at least five different interpretations of what was betting, what was gambling, was it skill, was it chance, what percentage and what should be the conveyance of value and things like that. In that sense, the new engines, I’m told, can not only provide precedents but given a judge with a 10 or 15 year track record, they can even tell you which way this judge is likely to rule on a particular case. The amount of processing is infinite. We have gone from a processing-limited society to a data-limited society. Theoretically, if you had infinite data, you could get the machines to learn everything. But we see limitations already. The engines in English are far superior to the ones in the regional languages. There are companies that try to fill that gap. I read somewhere about companies in Karnataka, who pay people in rural areas to read Kannada or Tamil or Telugu into a machine, so that they can build databases of audio in regional languages, for the machine to process. The ability of the machines to learn is still not quite as good as what a human can. So the volume of data required is going to be infinitely higher. But all that tells you is that you’re on a progression that is never going to reverse. It’s a secular trend. The more data there is, the smarter these learning models will become and the more difficult it will be for you to tell what is human and what machine output is.
Over time, the number of applications will get wider and wider and deeper and deeper. As the IT Minister, many of the major tech companies come to me. Their primary offerings are to automate the notion of a call center or dispute/complaint resolution desk. They believe that if we give them enough GOs and case studies, they can make the machine accept a query and give an answer, without the persons on the other side even knowing that it’s not a human being that answered them. Again, it’s easier to do in English and it’s harder to do in Tamil. We’re working hard on trying to bridge that gap.
A Minister’s job is safe, for now!
Many jobs are going to be affected by this transition—sometimes in bad ways, sometimes in good ways. I’m not personally worried, at least as long as I’m in public life, about being replaced by these machines. When I was Finance Minister, in two years, I processed about 7000 files. The chief secretary at the time, used to tell me that if I was away from the country for more than a week or 10 days at a time, then the system would come to a halt. Because all files go through the finance minister, some way or the other. I’ve had to change travel plans. Once I came from New York to Chennai for 24 hours, signed files, then went to London and attended the Oxford Union debate. I then flew from there to Melbourne. A human being is needed here to sign these files. Thinking about those files, on many of them I raised queries and that led to substantive change. It’s really hard for me to envision any machine being able to make that kind of automated decisions. So at least for now, I think the job of a Minister is fairly safe. But you never know. Even those jobs may be taken over by a machine sooner or later.
Panel Discussion:
CA Pattabhi Ram V: We are living in one of the most amazing periods of world history, a period which is pregnant with so many possibilities. Most of us who are in our careers still, are extremely fortunate. We see the magic of artificial intelligence happen right in front of our eyes. Once upon a time, it was just science fiction.
Mr P.V.Rajarajeswaran: Nowadays, government executives and policymakers are also talking about artificial intelligence and ChatGPT. It is said that there will be a big danger for the chartered accountancy profession if AI/ML and ChatGPT fully come into force. What is your view about the thinking of our government on artificial intelligence?
Ms Priya Bhansali: The Niti Aayog, in 2018, brought out a paper on artificial intelligence where they stressed upon responsible artificial intelligence. Later in 2021, they brought out a paper, again on the principles of responsible artificial intelligence. They stress upon seven characteristics of artificial intelligence that include integrity and risks. Positive human interaction is one of the principles of AI brought up by Niti Aayog. The government is not only thinking about artificial intelligence as such, but on how it can integrate it with positive human relationship. There has not been a lot of history on this, in terms of act or laws. Institutions like the ICMR have brought out guidelines on ethical biomedical research and healthcare. In 2019, Sebi brought out guidelines for inventorising the AI tools used in the capital market by brokers and market players.
The government, through the national education policy, is trying to include AI at school level education. The skill development program of the government also stresses upon it. They provide grants for institutions that skill people on AI. Europe has brought out an AI Act, which also talks about the risks. It means that programs which are risky will not be permitted to be used in those countries. The laws are strict there.
Mr P V Rajarajeswaran: There are both positives and negatives from ChatGPT. How does it help the education sector? Does AI disrupt the workplace?
Dr Anbu Thambi: AI is nothing but a super powerful word of automation. More than a decade ago, Microsoft introduced mail merge option. Even today, many of those in the offices don’t know how to use it. Then many advanced features were introduced in mail merge. For instance, as Chartered Accountants, you can use mail merge to send IT returns to thousands of companies, quite easily. The mail merge will pick up the return pertaining to a company, based on the email address and send that return to that company. It is a very powerful option. All these advanced levels came five years back.
The next super powerful thing which is coming now is that the tool can send a voicemail through WhatsApp, like the mail merge option. The messages can be customised for each person. Because of the cost factor, AI has not yet been commercialized. One fine day, it will become like a commodity and everybody will be able to afford these levels of automation. Today, in L&T, we conduct automated exams and interviews. There are some standard questions. As we are hiring freshers, we have to interview a lakh of people to take in a few candidates. The automated tools gauge exactly like a human. That power has already come into the workplace.
On the negative side, you can ChatGPT to even generate a PhD thesis but you can easily make out if the output is from ChatGPT. Now there are also tools to even convert the ChatGPT output into a human-like output. So, it is both a boon and bane to education sector.
Mr P V Rajarajeswaran: From the deliberations of the Court, I have a feeling that some of the judgments are also written taking cues from ChatGPT. How will ChatGPT help small and medium chartered accountancy firms to make audit programs and to prepare audit checklists?
Mr. Ganesh Balakrishnan: As Chartered Accountants, we need to have the basic knowledge of accounting and auditing. That’s where we need to start with. ChatGPT is an assistant. It will give you a list of information. You have to compare and check that information with the technical library or the information that you have. That’s number one.
As far as checklists are concerned, the ICAI themselves have come out with a lot of checklists and documentation. But when you go for a client meeting and you want all these information to be captured, ChatGPT can play a major role. You can use ChatGPT to generate many questions that the client may ask and to get probable answers for those questions. Many companies now use drone technologies to do inventory verifications. This is possible, only when you have a good system in place. When we use such technology, the error of judgment is very minimal.
Just as technology has evolved, humans have also evolved. Though there are tools available, it’s important for all of us to use them rightly. Learning is continuous. ChatGPT also learns from its experience. It is definitely a great tool, not only for the small and medium enterprises but for all the accounting firms. Some people use very high-level analytical tools. With these tools, the concept of sampling has gone. In fact, auditing standard itself should undergo a change. You can now do a complete analysis of the population itself. So why do you need sampling and materiality? There is now more responsibility on the auditors to report on fraud. Using artificial intelligence is the only way to make sure that we identify those red flags and address them.
Mr P V Rajarajeswaran: How does ChatGPT affect the younger generation? What is your advice to them?
Mr. Ganesh Balakrishnan: The younger generation wants to pick up things which are already available. If it is not available, you need to invest in that knowledge. The good part about them is that they are smart. They’re smarter than the last generation. The next generation will be much smarter than that. But what is important is that your fundamentals must be strong. That’s all you need. You can use a simulator and drive a car or an aeroplane. But when I give you the car to drive, would you drive it on the road or on the pedestrians or anybody else? That’s what becomes very important.
The younger generations are in the prime age of technology. But they need to understand the fundamentals of technology before they apply it. The youngsters have less patience. They want more money and want to convert everything in no time. They want to conquer everything. But at the same time, they must be more knowledgeable with their wisdom. After all, a technology gets obsolete. So, my suggestion to the young generation is: invest in technology, learn technology but be strong on your fundamentals.
