Choosing the Right AI Model is Key to Unlock the Technology's Fullest Potential

Shailesh is a seasoned technology professional with over two decades of experience in nice tech areas such as AI/ML, Deep Learning, Quantum AI, Computer Vision, Conversational Computing, and IoT, to name a few. He came on board Jio in 2018, prior to which he has handled key responsibilities across corporates such as OLA, Google, Microsoft, and Yahoo. Siliconindia recently got a chance to interact with Shailesh Kumar, Chief Data Scientist (CoE AI/ML), Jio, wherein he shared his insights various aspects to Generative AI, its implementation strategies and various other interesting aspects. Below are the excerpts from the exclusive interview  

What are your thoughts on the current Generative AI landscape in India?

There is no doubt that open AI has showcased its potential to all. Now, it is time for us to discover new, revolutionary use cases; and this is already happening on a daily basis. Since a generic English-only LLM will not serve everybody, the next step now is to build a specific language for each domains. This entire fine-tuning is the next wave of work that is happening for domains across languages. To summarize, the way we can expand English LLMs that are generic to domain-centric multilingual LLM is what the business fraternity is now focusing on. Another major challenge with LLMs is that we still must connect them structured data in order to derive personalized recommendations and answers. This will be the second area where will witness a lot of innovations happening in the near future. Thirdly, the system/application must be capable of proper reasoning before giving the answers by having that required level of compute reasoning capability.

Tell us about a few other technologies that businesses can leverage to enhance the effectiveness of Gen AI.

Although every cutting-edge technology that we have seen come-up over the past few years having been helping businesses in optimizing their operations, business leaders must be skillful in choosing the right kind of AI models for each use case and create the right AI architecture for the business. This is because today, we are living in a use case way of solving the problems. However, what is needed is the workflow way of solving the problem, which is identifying the workflows, knowing where AI & IT play a role, and having an integrated system in place where AI and IT team are working together to solve the problems for businesses. There is an immediate for having this shift in culture and mindset among businesses today. There is a severe dearth of AI architect across all business, and most of them just refer to online blogs which adopting AI. As a result, i strongly feel that we need to institutionalize the process of how to adopt AI into any vertical.

Throw some light on a few moral and ethical considerations to keep in mind while implementing Gen AI.

Firstly, the main purpose of using Gen AI is to derive accurate answers for critical reasons as education or healthcare. Thus, it is very important to make sure that the Gen AI integrated system is capable of deciding whether it knows the right answer and conveying the same to the user as well. Secondly, if the data that was used to train the AI models was gathered from one specific region, culture or demographic, the answer/insights might be applicable everywhere. For instance, the ideal sugar level for Indians might be different than that in the US. Hence, the adoption of Gen AI as per the data specific to each country and culture becomes very important. Thirdly, businesses must ensure the quality of data that is being used to train the AI systems and avoid bias in data to the fullest. Lastly, data creators should benefit and duly credited from the AI systems where there data is being used.

What does the future hold for in terms of tech advancements in India going into 2024?

There is a huge potential for India to create a new wave of information distribution in the country. This will no doubt uplift the literacy rate and tech skills among the general public. However, I personally am really excited to see the way Gen AI will drive decisions and policies across critical domains such as legal. Also, with the recent rollout of 5G, this will be the year of 5G and we have already started working on 6G as well. Thus, it is safe to say that use cases are now behind and technology is ahead. This will also raise the demand for 5G and its use cases, wherein I expect a lot of startups and other niche tech firms might come-up who the need for speeds in 5G levels to operate their daily functions. Going forward, while the telecommunication comes with higher bandwidth and becomes more reliable, the way people will utilize such high quality telecom solutions will see a lot of improvement.

Read More News :

Future Prospects for the Indian Pharma Sector - An Expert Overview

Manufacturing Industry in India - This Year's Story So Far & What to Expect from 2024