AI Standards Summit: Paving the Way for Inclusive and Sustainable AI Development
- 15 October 2024
By Celia Pizzuto
On World Standards Day, leading global experts gathered in New Delhi for the first part of the International AI Standards Summit, during the Fifth Global Standards Symposium (GSS-24). Setting the stage for the World Telecommunication Standardization Assembly (WTSA), the event marked an important moment in the ongoing global dialogue on the role of international standards in shaping AI.
Seizo Onoe, Director of the ITU Telecommunication Standardization Bureau, opened the discussions on AI standards, stating,
“ITU is proud to host the first edition of the International AI Standards Summit this week in New Delhi together with our longstanding partners IEC and ISO. We are committed to ensuring that standards support responsible AI governance. That is also the aim of our standards collaboration on AI watermarking and multimedia authenticity launched at this year’s AI for Good Global Summit. IEC, ISO, and ITU are now collaborating to develop an AI standards database, a key priority for the UN Secretary-General’s Envoy on Technology.”
The discussion then proceeded with a keynote address from Amandeep Singh Gill, the United Nations Secretary-General’s Envoy on Technology. In his opening remarks, he emphasized,
“We need a place, a marketplace […] for AI standards, for AI standards, where the community that’s working hard to develop AI standards fit for purpose for the coming AI rich economy can gather together, can deconflict their work, can identify gaps, can also build capacity of policymakers and others, to interface with these standards and reflect them in their own policy work.”
With this, he set the tone for a critical conversation on how AI standards can uphold human rights, promote safety, and foster sustainable development.
The AI Standards Summit’s first session brought together key stakeholders representing leading standards development organizations, including ISO, IEC, ICANN, ETSI, IEEE, IETF, and TSDSI, under the moderation of Bilel Jamoussi, Deputy Director of the International Telecommunication Union Telecommunication Standardization Bureau. The panel addressed three critical themes: the need for stakeholder dialogue, aligning AI standards with sustainable development goals, and identifying emerging trends in AI standardization.
The Importance of Stakeholder Dialogue in AI Standardization
The panel’s discussion started by focusing on the significance of dialogue between different stakeholders in addressing the complex challenges of AI standardization. Bilel Jamoussi set the context by emphasizing how knowledge of AI standards, data, and models is vital for guiding policy and broader decision-making to harness AI’s benefits while avoiding its risks.
Sergio Mujica, Secretary-General of ISO, highlighted the importance of inclusivity and transparency in the standards development process.
“Standards are the outcome of a conversation between experts that have been appointed by members, not by me in Geneva, but by the 170 members of ISO. And it’s also the outcome of a consensus that is being reached by them, working in a transparent and inclusive manner,” Mujica expressed.
He stressed that AI standards cannot be developed in isolation, as AI is a crowded field with overlapping standards, and global coordination is essential. Mujica also announced that the next AI Standards Summit will be held in December 2025 in Korea, further demonstrating the commitment to ongoing global collaboration.
Philippe Metzger, Secretary-General and CEO of IEC, built on this idea by emphasizing the complexity of today’s technological landscape and the need for coherence among standards bodies.
“With the increased inclusion of [socio-technical] dimensions, the coherence between different standards development organizations becomes even more important,” Metzger remarked.
He also highlighted the importance of ensuring that new standards address emerging challenges comprehensively, emphasizing that standards must not only be interoperable but also adaptable to rapid technological advances, ensuring inclusivity of diverse socio-economic needs. He noted that long-standing collaboration between IEC, ISO, and ITU is helping to address these challenges through joint initiatives.
Tripti Sinha, Chair of the ICANN Board of Directors, stressed the necessity of multi-stakeholder dialogues to prevent fragmentation in the AI landscape.
“Healthy dialogues prevent fragmentation. Without common standards there is a risk of a fragmented global AI landscape, with incompatible systems and approaches,” Sinha noted.
Sinha highlighted that multi-stakeholder conversations must be had on this topic to foster trust and responsible development of AI standards, emphasizing the importance of ensuring that all voices – including those from smaller or developing countries – are part of these discussions to create truly inclusive and effective standards. She stressed that it is only through ongoing engagement and cooperation across different sectors and regions that a unified and consistent AI ecosystem can be established, ultimately enabling shared benefits and preventing silos in technological advancements.
Issam Toufik, Chief Technology Officer at ETSI, echoed these thoughts, emphasizing the importance of avoiding fragmentation and fostering global standards.
“Fragmentation will be detrimental to innovation […]. We need to avoid this fragmentation and work as much as possible towards a global standard ” Toufik said.
He highlighted that collaboration across nations, sectors, and industries is crucial to achieving economies of scale and ensuring that the benefits of AI reach everyone, including developing countries.
Aligning AI Standards with Sustainable Development Goals
The discussion then centered around on how AI standards can be inclusive and aligned with the Sustainable Development Goals (SDGs), particularly in addressing the AI and digital divide. This conversation highlighted the necessity of a holistic approach to ensure that AI development benefits everyone, including marginalized and under-represented groups.
Karen McCabe, Senior Director of Technology Policy at IEEE Standards Association, emphasized the need for diverse and meaningful participation.
“We need to more effectively engage diverse stakeholders […] to enhance and explore new approaches to ensure that voices affected by AI systems are heard,” she stated.
McCabe underscored the importance of including non-technical attributes, such as gender-responsive practices and human rights considerations as a core component of standard-setting processes, ensuring AI technologies align with societal values.
Roman Danyliw, Chair of the Internet Engineering Task Force (IETF), highlighted the importance of balance.
“For all the promise and excitement AI-enabled systems are going to bring […], it’s imperative that [investments in AI] don’t come at the expense of all of the critical work already happening to address foundational issues that contribute to the digital divide, [such as] internet connectivity, affordability of that connectivity, and digital literacy,” he cautioned.
He also stressed that an open multi-stakeholder model is crucial for addressing the rapidly evolving challenges of AI while ensuring inclusivity.
Mohanram Natarajan, Director General of TSDSI, added that diverse participation from developing countries is key to bridging the digital divide.
“To make AI standards inclusive, we need to ensure diverse participation from all stakeholders, especially from under-represented regions such as developing countries,” he said.
Natarajan emphasized the role of capacity building and knowledge sharing to ensure that AI technologies are sustainable, inclusive, accessible and affordable for everyone.
Emerging Trends in AI Standardization
The panelists identified several key emerging trends in AI that would greatly benefit from international standards. One of the major trends discussed was AI regulation support. As more countries establish AI regulations, international standards can help ensure consistency and provide frameworks to comply with these regulations, including risk-based approaches that are proportional to the level of risk involved.
They also mentioned operationalizing AI principles like fairness, transparency, and accountability, stressing the need for standards that can translate these principles into practical tools across industries, thereby building trust in AI systems.
As the environmental impact of AI increases, particularly with the energy-intensive nature of large models, the speakers highlighted the growing need for standards that measure and mitigate the carbon footprint of AI technologies to promote sustainability.
The potential of multimodal AI, which processes multiple types of data like text, images, and audio, was discussed as well. Standards for interoperability and performance across these modalities are essential to ensure seamless functionality.
In healthcare, AI presents significant opportunities, especially in medical imaging and diagnostics. Standards are necessary to ensure safety, accuracy, and reliability in these critical areas.
Security was another major focus, with panelists noting that AI systems are both targets of attacks and potential security threats themselves. Developing standards to protect AI systems and leverage AI for enhancing security is crucial.
Other key areas included AI testing and quality assurance, the need for standards tailored to regulated industries such as healthcare and finance, and interoperability to enable AI systems from different vendors to work together. Finally, ethics and inclusivity were emphasized, with a call for standards that ensure fairness, accountability, and privacy, as well as inclusive governance of AI technologies.
A Commitment to Collaborative AI Standardization
The first session of the International AI Standards Summit at GSS-24 was a powerful reminder of the importance of global collaboration in AI standardization. From ensuring inclusivity and transparency to addressing sustainability and security, the discussions underscored the need for coordinated, multi-stakeholder approaches to developing AI standards that serve the broader interests of humanity. As Amandeep Singh Gill aptly noted, the journey towards an AI-rich economy requires a space where global experts can come together to build a trusted, sustainable, and inclusive future.
The conversation is far from over. The International AI Standards Summit will continue with further sessions on October 16, 17, and 18. These upcoming discussions promise to deepen the dialogue on how standards can shape the future of AI, ensuring that its benefits are shared across borders and communities.
