Latest news with #AccelAISummit

Business Standard

India's IT giants can't afford to sit out the AI race, says senior official

While startups and academic researchers have so far led India's push to build foundational artificial intelligence (AI) models, the country's legacy IT giants are unlikely to remain on the sidelines for long, according to Abhishek Singh, Additional Secretary at the Ministry of Electronics and Information Technology (MeitY). 'Given the capability that our IT industry has and what we are hearing about Infosys and TCs, all of them are working on AI-based applications,' Singh said at the Accel AI Summit, 2025 on Wednesday in Bengaluru. 'There will be a need for doing this and enhancing their capabilities.' Infosys and TCS have both begun ramping up investments in high-performance computing infrastructure, including Graphics Processing Units (GPUs), while simultaneously retraining thousands of employees in AI-related skills. Singh noted that these companies are already supporting global clients with AI-driven applications, and it's only a matter of time before they begin to roll out proprietary AI products and platforms. 'Otherwise for them to survive without AI, it will be very difficult,' he said. 'The whole world is going towards the adoption of AI.' He said whether it's software delivery, coding practices, SaaS implementation, or solutioning for enterprise clients, everything is being reshaped by AI. Singh believes the transformation underway across the global IT landscape will force India's large tech players to shift from primarily service-driven models to product-led AI innovation. The demand for intelligent applications and automation is only accelerating. And given their scale and customer base, India's top IT firms are well-positioned to deliver on that demand—if they move swiftly. While traditional IT services companies may not be leading the development of India's foundational AI models, they are quietly catalyzing a wave of innovation through internal transitions and ecosystem partnerships, according to Prashanth Prakash, founding partner of Accel India. Inside many large firms, there's an active rethink underway about what comes next in the age of AI. One notable trend, he noted, is senior leadership exiting established IT companies to launch or support new AI ventures. These leaders already understand enterprise pain points and high-impact use cases. They bring domain expertise and built-in demand. 'They're now partnering with VCs to build startups from the ground up,' he said. Another model emerging is through Global Capability Centers (GCCs) based in India. These units are increasingly approaching startups with concrete enterprise use cases and positioning themselves as early partners. They are asking the startups to take their workflows and reimagine them using AI agents or automation. This convergence of talent migration, early-stage venture activity, and enterprise-driven use case development, signals a deeper transformation within India's services ecosystem. 'I think we're seeing multiple ways in which there will be disruption within the context of the Indian services ecosystem,' said Prakash. AI regulations As artificial intelligence applications proliferate across sectors, the Indian government is also taking a pragmatic approach to regulation—aiming to ensure compliance with existing legal frameworks while avoiding heavy-handed restrictions that could stifle innovation, according to Singh of MeitY. 'Regulation will primarily be to ensure that any AI application that is developed is compliant with the legal framework and the laws as they stand,' Singh said. 'But we are not inclined to have something similar to the European Union's approach.' Instead, Singh explained, the government will rely on sector-specific and harm-based safeguards anchored in current statutes—such as the Digital Personal Data Protection Act (DPDP), the Information Technology Act, and the Bharatiya Nyaya Sanhita, which recently replaced the Indian Penal Code. Citing examples, Singh emphasised that generative AI models must be sensitive to local legal contexts. He gave the example of pre-natal sex determination, which is strictly prohibited in India. Yet, an AI model trained on global datasets might identify the sex of a fetus from an ultrasound image, violating Indian law. He said developers will need to ensure such models are adapted to local legal and ethical norms. The focus, Singh added, will be on preventing harm—such as the spread of deepfakes, misinformation, or content that could incite violence or infringe on individual rights. He was of the view that if an application risks violating the law or causing harm to individuals or communities, it will be regulated. However, Singh was clear that regulation would not come at the cost of technological progress. 'We are more inclined towards promoting AI application development rather than restricting it.'