Activities
Divisions
Programs
Activities
Divisions
Programs
FP Special: Amit Luthra, Managing Director (ISG) of Lenovo, discusses the importance of Autonomous Data for Autonomous AI
Approximately 93% of Indian corporations intend to put money into AI by 2024, focusing not only on GenAI and LLMS, but also on AI inferencing patterns. In this context, creating Autonomous AI is essential, but achieving this without Autonomous Data will be highly challenging.
Even though artificial intelligence has shown remarkable progress up until now, we're only observing the beginning stages. Although our main attention has been on generative AI and LLMs, there are numerous advancements happening in other areas, and they're evolving at a much quicker pace.
For instance, 93% of Indian corporations intend to put money into AI by 2024. Their focus isn't solely on GenAI and LLMS, but also on AI inferencing systems that assist them in examining and understanding scenarios, and predicting necessary actions. While GenAI has elicited both amazement and fear, it's analytical AI, automation, edge-computing and AI inference that represent the future for Indian businesses.
We had a conversation with Amit Luthra, who is the Managing Director for Infrastructure Solutions Group at Lenovo India, during the reveal of their 'CIO Playbook 2024 – It's all about Smarter AI' report. We talked about the future of AI in India.
We cover everything from the concept of data autonomy to the moral structures surrounding AI. We also debunk the misconception of AI "taking" our jobs and discuss how Chief Information Officers intend to utilize AI, not just because it's a trending term, but due to its capacity to make a significant difference. Here are the condensed highlights of our discussion:
What changes have occurred in India's technological requirements over time, particularly in relation to AI? India's digital transformation was primarily based on four aspects – Social Media, Mobility, Analytics, and Cloud Computing. In the case of Analytics, there was always an element of AI. Most business applications have attempted to incorporate AI into their operations. However, two significant events have brought about a shift in this scenario.
Initially, the arrival of the Internet of Things resulted in a surge of data creation from various sensors. Businesses acknowledged that they couldn't afford to collect the data first and then examine it retrospectively to achieve their objectives. For instance, if a manufacturing plant aimed to enhance its production quality, it was crucial to scrutinise the data as it was being collected, particularly if the data was time-dependent. This situation amplified the demand for Edge Computing.
The arrival of GenAI and the creation of extensive LLMs would be the next step. It requires advanced machines with significant computational strength to construct such large LLMs.
Nonetheless, I believe that a substantial portion of the current investments in India is centered around developing LLMs. However, we will need to shift our attention towards boosting investments in AI inferencing in the future. The growth of inferencing will likely occur at a considerably rapid rate. This will involve a combination of CPU-GPU workloads, instead of solely depending on GPU-based systems.
What are the primary obstacles that businesses, particularly startups, face when implementing AI? AI, specifically generative AI, requires a well-organized structure. The main hurdle is to establish efficient data management and maintain data accuracy. If your AI models are trained on poor quality data, they will produce poor quality results. Even if you develop an excellent LLM that learns rapidly and possesses incredible algorithms, it will cause problems if the data used for training is incorrect, incomplete, outdated, or of poor quality.
It's also vital to create an analytical structure that aligns with particular objectives. Businesses must understand the rationale behind their need for an AI system and their intended outcomes. They should question, "What is my goal with the collected data?"
Many AI initiatives in numerous organizations falter due to a lack of precise understanding of their AI goals. While certain companies aim to enhance customer happiness through AI, others may seek to boost their production levels.
Some might wish to establish cleanliness routines to boost efficiency, while others might be more interested in preventing pilferage and examining fraudulent activities. Companies must be certain about their ultimate goal and what they intend to accomplish. Once this is established, the appropriate technology can be aligned accordingly.
Next, considerations must be given to both the security and ethical aspects, both of which are equally significant. When handling data, it can either elevate or devastate you. It's crucial for businesses to contemplate how to handle the ethical side of things and determine the boundaries. There is also a need to manage delicate data correctly, implementing an appropriate level of care and a robust security system.
The CIO Playbook 2024 indicates that nearly half of all CIOs face difficulties when trying to fill AI-related positions. Despite India having the highest number of engineers and developers, it seems there may be a gap between the industry's needs and the existing abilities of our engineers. AI has been a part of the engineering curriculum for approximately 25 years. However, applying what's learned in an academic environment to the real-world corporate setting is a separate challenge. We can't assume that a recent graduate can instantly demonstrate their competence in AI work.
The issue isn't a lack of resources, but rather creating a supportive structure for such resources.
This can be especially difficult due to the rapid pace of change. What was deemed essential last year might be irrelevant this year. Business owners may now be questioning how they can constantly guide and educate their teams, or how they can consistently implement a learning structure.
The problem of retention also arises. The act of poaching is a grave concern, particularly in relation to AI. In this context, companies must consistently allocate funding to their resources and must also possess the determination to retain these resources.
There's considerable discussion happening about Sovereign AI and Data Sovereignty. In light of the current development of LLMs, what hurdles do you anticipate in creating sovereignty? Data sovereignty refers to being aware of your data's location and making sure that the data remains within your country's geographical borders. Usually, data might be kept on the cloud, but the physical storage of the cloud could be in a different country.
For instance, numerous multinational corporations previously maintained unified data centres situated anywhere globally to enhance and streamline their processes. In such scenarios, sovereignty isn't possible since the data – potentially the private details of Indian nationals – is stored beyond the country's borders.
In this context, the importance of data sovereignty legislation is highlighted as it mandates that personal data of Indian citizens, for instance, must remain within the country.
After confirming this, the security and analytics framework along with the government step in. It is their role to guarantee that crucial data isn't exposed, and it's managed correctly and stored according to set guidelines.
So, how do businesses in India utilize this data and create an analytics structure around it? It's crucial to understand what your goal is and what you're aiming to accomplish. Each sector and application is unique.
In that hypothetical situation, what sort of LLMs or apps might we anticipate being created? Here's an illustration. If I have to contact the customer support team of a service I'm using, I'd rather give them a call than interact with a chatbot on an app, primarily due to the lackluster user experience. However, we're now witnessing the emergence of AI-powered voice assistants acting as customer service representatives, boasting a fully voice-enabled interface.
Consider the retail industry as another case. In the near future, we'll see digital representations replacing the average salesperson, designed to answer your unique questions tailored to your precise needs. Imagine walking into a car showroom and inquiring about the specifics of a new vehicle. A less seasoned salesperson might muddle the information, leaving the customer confused. Furthermore, they might struggle to respond if you pose a highly specific, but infrequently asked, question about the product.
Picture the same situation, but this time with a virtual character who is almost identical to a human, and is fully knowledgeable about the product they are responsible for. They can respond to all your inquiries after only a single training session.
Additionally, the company does not have to invest significant time in retraining employees whenever there are updates or changes. In this scenario, the only requirement is to make sure that the language models, regardless of their size, are continually learning and have access to the most recent data sets. Think about how distinct the previous versions of ChatGPT are compared to the versions powered by GPT4.5.
After the creation of the LLM, it's crucial to ensure its portability, modernity, and most significantly, its simplicity in deployment. This is where Lenovo steps in, assisting organizations with a spectrum of solutions that span from handheld devices to cloud-based systems. This is because AI is not confined to a single LLM; it pervades the whole ecosystem.
The Indian government has not yet established rules for AI within the country. What aspects would you like to be included in these regulations from a business point of view? Moreover, how do you think the Digital Personal Data Protection Act of 2023 will influence the progress of AI in India? It is essential to establish a robust framework for security and ethics. How can we protect individuals from deepfakes that are already in circulation? The conversation around security and ethics will be crucial. Once the landscape is clear, companies constructing Large Language Models (LLMs) will have a better understanding of the boundaries.
Corporate executives and business leaders, regardless of their proficiency in AI or other technologies, are eager to integrate AI into their operations. How can CIOs effectively communicate with them to guarantee that AI is correctly and appropriately utilized? The desires of the business sector and the technology sector often differ, as our research for the CIO Playbook 2024 has shown. Top executives are inclined towards generative AI, possibly because it is currently a popular term.
In 2023, Generative AI wasn't a significant topic, but now it's a key priority since it's gained considerable attention. Last year, the primary focus was on enhancing customer experience and satisfaction rates. But now, these aspects have taken a backseat as businesses realize that while attracting new customers can be costly, keeping existing ones is comparatively simpler.
Likewise, the growth of revenue and profit, which has always been seen by all companies as their top priority, has now dropped to third place. Sustainability, which was never a main concern, has now risen to the fifth most important priority due to ethical considerations.
Let's now shift our focus to the preferences of the CIO or tech department in a company. CIOs continue to hold the view that automation should be the top priority. By placing emphasis on automation, numerous aspects can be managed efficiently. Thus, AI, particularly generative AI, is considered subsequent to establishing automation.
It's necessary to establish a specialized infrastructure designed to handle the intensive tasks of AI. Furthermore, it's essential to embed security and ethical considerations into the heart of your AI. As previously mentioned, this will serve as a structural guide. AI doesn't function autonomously, it must be directed towards a goal. Incorporating AI is akin to formulating a strategy, involving various elements and requiring an organizational structure. This is something we've always championed and we're pleased to see this sentiment affirmed in our survey, as we revisit discussions with CIOs.
Which aspects should the government prioritize when incorporating AI into their administrative tasks? How can AI improve the efficiency of the government? The government has established its own perspective on AI. As citizens who have voted for our government, we should support their perspective. When government officials and authorities engage with industry experts like ourselves in discussions, they gather various insights and opinions on different topics. They also encourage us to contribute our views and suggest optimal strategies.
I believe that when the government develops a vision for AI and establishes a structure such as data sovereignty, everyone will need to comply with it. Additionally, our Prime Minister, Shri Narendra Modi, has spoken about 'AI for All'. We at Lenovo have been promoting the concept of 'AI for All' for several years. We also hold the view that when AI is available to everyone, this vision will be greatly reinforced.
Ultimately, everyone is wondering: How will AI alter the employment and job landscape as we know it today? AI should be viewed as another technological device or suite of utilities that will enhance human performance. Take this for example – soldiers participated in World War I, World War II and all other wars we've waged. During World War I, soldiers were equipped with .303 rifles, but today, they use LMGs and MMGs. Despite the transformation of tools and the immense progression in technology, soldiers are still required.
AI is simply another resource similar to firearms, another facilitator that will allow them to perform more efficiently and manage advanced equipment.
Indeed, AI will alter the characteristics and possibly the standard of jobs. However, it won't eliminate jobs or render them obsolete. While we may not require certain types of roles, new opportunities will simultaneously emerge.
I'm of the opinion that AI will lead to an enhancement in salaries, as it necessitates the acquisition of new and superior skills. Individuals must commit to self-improvement or risk becoming irrelevant. Consider the transformation in the IT industry. Two decades ago, I regularly queried job applicants about their knowledge of SAP sizing and exchange sizing during interviews. However, such questions are no longer relevant.
If a person claims to be highly skilled in PowerPoint on their resumes but lacks the knowledge of integrating it with AI, they are not adequately qualified. AI is a crucial tool for our transformation. Ultimately, it's about staying pertinent, securing employment, and maintaining industry relevance. For this, self-investment is necessary.
Look for us on YouTube
Headline Shows
Connected Articles
NVIDIA's Jensen Huang claims AI hallucinations can be tackled, predicts artificial general intelligence within 5 years
Apple finally unveils MM1, its AI model capable of generating text and images
Microsoft brings on board Mustafa Suleyman, DeepMind cofounder, to head its new AI team for consumers
Samsung and Rebellions, South Korean chip producers, aim to outperform NVIDIA
AI hallucinations can be addressed, artificial general intelligence anticipated in about 5 years: NVIDIA’s Jensen Huang
Apple successfully introduces MM1, its AI model designed for text and image creation
Microsoft recruits Mustafa Suleyman, DeepMind's cofounder, to steer its fresh AI team for consumers
South Korean chip makers Samsung and Rebellions have set their sights on surpassing NVIDIA
Available on YouTube
Firstpost retains all rights, as protected by copyright law, in 202
+ There are no comments
Add yours