Independent Advisor

1) Achieve multistakeholder cooperation though leveraging the existing forums (specialised task force to reduce overlaps) 2) Consensus on a prioritized 'Global AI Governance Roadmap' including but not limited to standardised risk frameworks. 3) Resource mobilization to prioritize capacity building, development of interoperable standards, especially audit tools for emerging markets etc. 4) Defined and agreed red lines for high-risk AI like autonomous weapons, deepfakes, bioweapons etc. Facilitation/creation of oversight bodies to closely monitor high risk/impact innovations. 5) Concrete pledges from diverse actors: Member States on national reporting; industry on transparency benchmarks; civil society on equity monitoring, rights protection and independent assurance. 6) Charter for ongoing engagement - Metrics to track voluntary commitments by states, onboarding of strategic partners (AI Ethics Hub/Safety Institutes) and independent experts.

• Safe, secure and trustworthy AI
• AI capacity-building
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Transparency, accountability, and human oversight

I selected the broadest options which covers the entire list with all its critical elements. Ex- Safe & Trustworthy AI could mean systems that protect human rights. It could also include standard and open data, models and software (where needed) to make it equitable for technology development and adoption. Similarly, technical implications of AI could include interoperability of governance approaches. As an AI Ethics advisor focused on audits and governance, Safe and Secure AI is foundational for risk frameworks specially in high-risk applications like deepfakes addressing urgent safety gaps. Capacity-building is non-negotiable, especially for Global South regions needing practical audit tools, standards, and skills to contribute and have a voice instead of being a passive observer. This is another area where I am deeply involved as a "ForHumanity" Auditor (FHCA) and Contributor. The social, economic, ethical, cultural, linguistic and technical implications of AI cover the rights/societal angle as well as technical aspects holistically as it captures equity, biases, linguistic fairness, and economic disruption that narrower human rights framing might miss. As a contributor and member at "AI & Equality" and EU projects - DIVERSIFAIR, BIAS, Mammoth and Aequitas my work at the intersection of socio-economic-legal lens for use cases like automated decision making in critical applications (hiring, healthcare, public services) help me prioritise fair and responsible usage of AI and create awareness. Finally, transparency, accountability, and human oversight provide the enforcement backbone through benchmarks, reporting, and independent checks extremely critical as we move from principles to practice through the open and crowdsourced, binary audit criteria that we develop, contribute, leverage at ForHumanity. Prioritising these creates actionable clusters for the roadmap and positions the Dialogue as a coordination hub responsible for driving the much-needed change and the much-needed pace.

Although AI's energy and environmental footprint cuts across safety, capacity-building, and social implications it lacks explicit focus, which would be good to have.

Challenges: No standalone AI law means fragmented rules (IT Act 2000, DPDPA 2023), leading to ambiguity on liability, algorithmic bias in diverse populations, and weak oversight. Gaps in skills and infrastructure (e.g., regulatory sandboxes to test high-risk applications). Availability of trust and transparent benchmarks. Opportunities: IndiaAI Mission's Safe & Trusted pillar enables ethics hubs/audits. Capacity-building via hybrid channels (outreach to Tier 2/3 cities; covering unemployed youth). Sovereign data/models addressing social implications (linguistic biases). Global Dialogue inputs could fast-track IndiaAI Safety Institute and/or sector specific market watchers, boosting exports in trustworthy AI services plus DPI/interoperable standards for large-scale initiatives.

1)It should consolidate fragmented efforts by mapping overlaps into a "Global AI Governance Roadmap," prioritising standardised risk frameworks and interoperability standards, turning repeated themes into phased action -> capacity-building, red lines for high-risk AI, testing and validation facilities. 2)Foster multistakeholder pledges ensuring inputs from diverse actors- industry, academics and research institutions, civil society, with focused efforts to include representatives from Global South. 3)Address energy and environmental footprints through evidence from the AI Scientific Panel 4) Resource mobilisation for risk mitigation in critical areas - border security, data and technology transfers etc.

Along with the existing partnership channels through GPAI/OECD, IndiaAI MIssion, UN Global Digital Compact , there should be provisions for engaging independent experts throughout the process for unbiased audits and evidence.

Answered earlier

Women in the Global South

Hybrid works best. Great to have an opportunity for written submissions which are as detailed as this.

Other than references like EU AI Act, NIST RMF, ISO 42001, prEN 18286, I have high expectations from the regulatory sandboxes - Singapore/UK, EU (Interoperability Act) and the National Digital Twin Initiative, 'Sangam' India as they are strongly positioned to guide the policy, law and technical standards in an ongoing manner. Also, forums like ForHumanity, AI & Equality, Women@theTable, AI+ Alliance, Eticas, SciencePo, Turing are working relentlessly on capability building, independent audits, evidence-based policymaking and creating awareness.