Clicky

Home » Latest News » Tech » TCS is building its own ChatGPT-like generative AI tech

TCS is building its own ChatGPT-like generative AI tech

By

Ankita

| Updated on:

Tata Consultancy Services (TCS) is looking forward to building its own version of ChatGPT for enterprise code generation. As of now, the project is in the initial stage and is harnessing extensive internal resources, codes, and data. The company will develop the solution using in-house algorithms which will use large language modeling functions. 

TCS is building its own ChatGPT-like generative AI tech

Key Points: 

  • TCS is building its own version of ChatGPT for enterprise code generation. 
  • The company is conceptualizing a role within the organization known as “Prompt Engineer.”

TCS building its own ChatGPT for enterprising code generation 

Chief Operating Officer of Tata Consultancy Services N Ganapathy Subramaniam stated the company is creating its own ChatGPT which will be used for enterprising code generation. 

Right now the project is in the initial stage and is looking forward to harnessing the vast internal code, resources, and data. The solution will be developed using in-house algorithms which utilize large language modeling functions that are similar to generative AI chatbots such as OpenAI’s ChatGPT but instead for enterprise code generation. 

Subramaniam stated, “The way we are looking at it, generative AI utilizes past code, experience, and data to comprehend,” In addition, TCS has been in the business for several years, therefore the knowledge can be used as a base. So, if the technology develops codes that have been taught algorithms using TCS proprietary data. In that case, the result is going to be something for which I am willing to license. 

In order to build proprietary generative AI algorithms, there are various use cases that need to be worked on by TCS to build code and the process has to mature later. Subramaniam also believes fine-tuning is still required for enterprise usage of AI solutions when it comes to ownership as well as the IP of the end products. Therefore, enterprise solutions will see demand. 

Subramaniam, who will retire next year in May, expects the Mumbai-based software firm to launch some of these solutions before the end of his tenure. The main aim behind generating these models is to create such codes which can easily be deployed. 

He added, if TCS wants to develop secure and contextual generative AI solutions to speed up clients’ projects, then it cannot simply depend on large language tools from various providers but instead should have its own solution. 

“Clients are willing to utilize a generative AI, said Subramaniam. However, as a supplier, I am obliged to provide them with something which is free of any infringement and contains the complete ownership of the intellectual property.” 

In addition, the industry needs to also wait for regulations governing these solutions to mature further to release the full capability of the technology. 

Development of new roles for “Prompt Engineers” 

According to Subramaniam, there is a massive opportunity in developing new roles for “prompt engineers” in the organization who can effectively communicate contextual questions to large language models for the requirement of enterprises. 

Therefore, one thing that we kept on listening to in the market was the way to ask questions on the AI platforms, which is extremely crucial, right? 

So, we are conceptualizing a role within the company known as “Prompt Engineers.” He added, “We are yet to get authorization on it, however, we are aware that the role will be as critical as having engineering, analytical or mathematical skills.” 

TCS saw an opportunity to collaborate with generative AI offerings using MasterCraft, its low-code software development platform. 

This will be quite similar to CoPilot, a solution to auto-generate code, which was released by Microsoft-owned GitHub and OpenAI to speed up the writing process of the software program. 

More than 100 clients are utilizing the Mastercraft tool. “As of now, we are yet to add a ChatGPT-like answer for data pattern intelligence.” We haven’t yet put a ChatGPT-like solution for data pattern intelligence into it. However, when we eventually do it, our vision will be an intelligent low-code platform for MasterCraft. 

Subramanian said, there has been a variety of several developed by TCS that utilize generative AI solutions while functioning with global partners.

Leave a Comment