Towards shared AI safeguards: Key outcomes of Indo-German dialogue at the Trust and Safety Festival India 2026

The panel consisted of Ankit Bose, Head of Nasscom AI, representing the industry perspective; Vaibhav Ppulekar, Assistant Vice President Cyber Security and Global Business Line Manager at TÜV SÜD South Asia, representing the technical perspective; and Tarunima Prabhakar, Founder and Research Lead at Tattle Civic Tech, representing the research and civil society perspective.

Not “whether” but “how” to implement open-source AI

Julius Johann Klüpfel, Economic Advisor at the German Embassy New Delhi, opened the session emphasising that the discussion should not be whether AI should be open, but how openness can be made trustworthy through standards and shared responsibility. He called for the discussion to move beyond the abstract debate and focus on practical collaboration by businesses, technical communities and civil society.

The panellists discussed how open-source AI can broaden access, reduce dependence on a few providers and support AI that reflects local contexts. They also raised concerns about misuse, bias and accountability. The Indo-German AI Pact, launched in February 2026, prioritises trustworthy and ethical AI and offers a framework for turning these discussions into cooperation.

Making AI safer through accountability and governance

The speakers highlighted that sharing an AI model’s underlying components does not automatically make it transparent or safe. Instead, safety measures must cover the entire system, and those responsible must be held accountable for harmful AI models, services and their use. Safeguards should be built in from the start, adapted when the system is deployed, and regularly tested and updated throughout its use.

Technical safeguards must be backed by effective governance. Clear policies, traceability and political guidance help build trust in regulation and are essential for the widespread adoption of AI safety measures.

Indo-German cooperation on trustworthy AI

The panellists put several concrete ideas on the table for Indo-German cooperation on safe and trustworthy AI:

• Align AI safety requirements and risk assessment, prioritising in two or three sectors, such as manufacturing, energy and healthcare
• Connect the open-source AI ecosystems of the two economies more strongly to facilitate exchange on responsible AI development and adoption
• Promote uptake of joint safety measures by Indian and German companies
• Build stronger links between research, industry and civil society of the two countries

The discussion showed that trustworthy open-source AI depends on safeguards at the system level and across the lifecycle. It also needs clear responsibilities, testing that reflects local languages and contexts, and practical tools. The Indo-German Digital Dialogue gives both countries a platform to take the ideas for cooperation ideas forward.