Nasscom, the trade association for the IT and business process management industry, has released a comprehensive set of responsible artificial intelligence (AI) guidelines for generative AI (GenAI). These guidelines aim to establish normative obligations for researchers, developers, and users of GenAI models and applications, ensuring the responsible adoption of this groundbreaking technology.
Extensive consultations with AI experts, researchers, and practitioners from academia, industry, and civil society were conducted to develop these guidelines, which were released in the first week of June. The goal is to create common standards and protocols that govern the development and use of GenAI. Nasscom said that it recognises the rapid advancements in the field and the potential societal risks associated with unregulated deployment, emphasising the urgent need for stakeholder cooperation in developing responsible approaches.
The Nasscom guidelines came a day beforeSam Altman, CEO of OpenAI, met Prime Minister Narendra Modi and discussed the potential of AI and the need for global regulation. According to Nasscom, the guidelines could serve as a fundamental framework for building stakeholder consensus on the core normative obligations in GenAI research, development, and use.
While they do not function as an operational manual, they are expected to drive the development of frameworks, standards, and best practices that mitigate potential harms.
Nasscom plans to release future editions of the guidelines to incorporate revisions and updates reflecting the evolving technological and regulatory landscapes.
Focus on GenAI
The guidelines specifically focus on GenAI, encompassing AI technologies capable of generating various forms of content. They apply to stakeholders across organisational affiliations, sizes, and geographical locations involved in GenAI research, development, and use.
Key principles outlined in the guidelines include adopting a human-centric approach, promoting transparency and explainability, ensuring accountability, prioritising privacy and data governance, advocating for fairness and equity, emphasising robustness and safety, and fostering collaboration and cooperation. These principles underscore the importance of designing and deploying GenAI in a manner that respects human values, ensures fairness and accessibility, and upholds privacy rights and data protection.
To ensure compliance with responsible AI practices, the guidelines recommend several implementation measures. They include conducting ethical reviews, empowering users with information and control, adhering to responsible data practices, mitigating biases, conducting robust validation, continuously monitoring deployed GenAI systems, and promoting public accountability and auditability.
Nasscom said that by adhering to these guidelines, stakeholders can collectively build a foundation of ethical and accountable GenAI practices that prioritise human well-being and societal values.
These guidelines come at a time when the international community is also recognising the significance of AI governance. Leaders of the G7 nations have called for the development and adoption of international technical standards for trustworthy AI. This global focus on AI regulation highlights the need for responsible and accountable practices to ensure the accurate, reliable, safe, and non-discriminatory use of AI systems.
Anivar A. Aravind, a public interest technologist based in Bengaluru, welcomed Nasscom’s guidelines for ethical AI use. He said: “It’s a good step for the industry to build its own guiding principles.” However, he added, more attention should be given to the rollout side, given the limited data sets and higher potential for bias in the country.
Meanwhile, a study by FICCI and Access Partnership, published on June 8, said that generative AI has the potential to increase the country’s productivity by $621 billion. It said AI can bring significant positive changes to the economy and contribute to its growth and development on a large scale.
With inputs from Saatvika Radhakrishna.