Generative AI isn't a replacement for traditional tech, it's an augmenter, says Google Cloud's Subram Natarajan

Generative artificial intelligence (GenAI), one of the hottest tech buzzwords now, should not be seen as a replacement for legacy technology but, instead, serve as an augmenter for traditional processes, said a top Google executive at TechSparks 2023.

Speaking to Shradha Sharma in a fireside chat, Subram Natarajan, Director of Customer Engineering at Google Cloud India, opined that "it is probably wrong" to view GenAI as a substitute for existing technology.

When it comes to the implementation of GenAI, organisations should fully think through its use cases as it can impact the internal functions of companies and how they deliver services to the end customer.

"There is also an influence on the culture of the organisation,” he added.

Natarajan also spoke about the impact of GenAI on content creation and how it could be used to efficiently drive home a point. He also warned of its pitfalls and said it is necessary to ensure that the content generated by this new-age technology follows all the necessary guardrails.

The Google executive highlighted that content discovery is another key area where users can leverage GenAI.

For instance, if you have a huge data repository with several gigabytes of data, the technlogy can help you parse through that mountain of data to fulfil your specific content needs. Besides content discovery, GenAI could also be used to generate personalised content that is tailored for your audience and engagement targets, he said.

Responsible AI and safety

On establishing guardrails for GenAI tools, Natarajan reckons it is essential that the data used to train these models follow corporate governance rules as well as other laws around copyright and anti-competition.

“The ability to create content using GenAI also comes with an added responsibility. Where am I getting the necessary knowledge to generate this content? Am I taking into consideration the licensing regulations that is allowed by companies? These are all big organisational guardrails that are necessary to be followed."

In Google, any image that is generated from text prompts is watermarked automatically and is indelible, thereby ensuring complete transparency, he pointed out.

Speaking about the rapid evolution of technology, Natarajan threw light on how Google now undertakes feature release cycles once a week compared to once a quarter earlier.

"To keep pace, you need a different kind of culture with shifts in GenAI. The right set of people, specific use cases, and right culture with respect to learning are essential to implement Gen AI," he stated.

In June, Google Cloud had announced that it would partner with Genpact to help businesses accelerate their AI adoption and strategies, including taking advantage of GenAI to derive actionable business insights that impact decision-making.