To accelerate and scale the impact of data and AI across enterprises, organisations must be hybrid by design : Siddhesh Naik, Country Leader, Data, AI & Automation software, Technology Sales, IBM India & South Asia
In an exclusive conversation with Express Computer, Siddhesh Naik, Country Leader- Data, AI, & Automation Software, IBM India & South Asia, highlights the significance of hybrid and multi-cloud environments in enterprise strategies, particularly in the context of GenAI solutions. He emphasises the need for a hybrid approach, identifying how cloud adoption is changing, accelerating the impact of data and AI in businesses, and enhancing business outcomes.
Some edited excerpts from the interview:
Given the growing emphasis on hybrid and multi-cloud environments, how is IBM incorporating these approaches in the context of GenAI solutions?
Enterprise-grade AI, including generative AI, requires a highly sustainable, compute-and-data intensive distributed infrastructure. Since AI workloads will likely form the backbone of mission-critical workloads in the future and manage the most-trusted datasets, the system infrastructure must be trustworthy and resilient by design. Generative AI will be multimodal and function as part of an ecosystem, with clients using a combination of third-party models, their own proprietary models, and open-source models. This hybrid approach to AI is similar to the hybrid approach to the cloud. To accelerate and scale the impact of data and AI across the enterprise—and ultimately improve business outcomes – organisations must be hybrid by design.
How do you envision the integration of GenAI with quantum technologies, if at all?
We are pioneering the use of generative AI for quantum code programming through watsonx. This aims to boost efficiency in quantum development, showing the potential synergy between generative AI and quantum technologies, shaping the future of AI applications. We are working towards integrating generative AI available through watsonx to help automate the development of quantum code for Qiskit. This will be achieved through the finetuning of the IBM Granite model series.
As GenAI evolves, how are you ensuring robust security and privacy measures to protect sensitive data and prevent potential risks?
As AI (including generative AI) adoption scales and innovations evolve, so will the security guidance, as is the case for any technology that’s used by enterprises. Given the recent advancements made in cybersecurity, we already have substantial tools, protocols, and strategies available to us for the secure deployment of AI. Securing AI requires building governance into the AI lifecycle with automated tools that direct, manage, and monitor an organisation’s AI activities.
When creating a robust security architecture, there are three key areas organisations must focus on. The first is to secure the data as it is the primary attack target in the data collection phase. As enterprises centralise and collate massive amounts of data to maximise generative AI’s value, that centralisation will also introduce new risks. Next is to secure the model by assessing and managing security risks that are present during the model development and training phases of the AI pipeline. Finally, secure the usage, as the model inference and live use phases of the AI pipeline are ripe for prompt injection attacks, model theft, and model denial of service attacks.
What future trends do you foresee in the realm of GenAI, and how do you predict its trajectory in the coming years?
Generative AI is poised for widespread adoption, and its use cases, especially in the B2B domain, are set to drive far better efficiencies and cost optimisation – provided we use AI responsibly. Our recent IBM AI Adoption Index 2023 found that six in ten IT professionals at Indian enterprises report that their company is actively implementing generative AI, and another 34 percent are exploring it. This points towards the pace of transformation in this space being explosive in the next few years. Hence, there will be a need to employ software automation to strengthen one’s ability to mitigate risks, manage regulatory requirements, and address ethical concerns for both generative AI and machine learning (ML) models.
We will also see roles like Chief AI Ethics Officer becoming critical for enterprises to ensure responsible design and deployment. Reskilling will play a very important role. Our report found that only 46 percent of the companies surveyed in India are currently training or reskilling employees to work together with new automation and AI tools, but it still leaves room for a lot more to be done.
