SingularityNET Decentralization Program Lead Esther Galfalvi shares her perspective on the importance of decentralization and global collaboration in the realm of AI.

Dear Singularitarians,

This is an insightful article by Esther Galfalvi, the Decentralization Lead at SingularityNET, focusing on the pivotal role of decentralization in achieving Beneficial Artificial General Intelligence (BGI). During the BGI Summit, Galfalvi participated in and moderated panels that delved into the importance of decentralization, global collaboration, and diversity in the realm of AI. She emphasizes how a diverse and crowd-governed approach is crucial for creating AGI that truly benefits everyone, highlighting the current dominance of a narrow demographic in AGI development.

“During the Beneficial AGI Summit 2024, I had the pleasure of appearing on the decentralization panel. It was wonderful to speak with such knowledgeable co-panelists and hear their unique perspectives on decentralization.

And of course, the panel was beautifully moderated by Dr. Mihaela Ulieru, whose enthusiasm and deep knowledge of the subtleties of decentralization made the whole experience a joy.

I was also pleased to moderate a panel on global collaboration and diversity, where we explored how countries worldwide are benefiting from AI, and how we could include more cultures in creating AI and AGI.

These experiences made me reflect on how diversity and crowd-governance is really at the forefront of how we will achieve beneficial AGI that exists as a balanced creation of all peoples, not just reflecting a few demographics.”

The panel led by Dr. Esther Galfalvi also delved into other interesting topics. One of those was living with robots, and the fact that doing so can demystify and remove fear of technology, as well as naturally turn into dependency and comfort. It also explored topics such as working with children to introduce them to coding and AI concepts, the vision of democratizing knowledge and bridging equality gaps, as well as the empowerment that Web3 gives individuals worldwide to access decentralized finance on par with the traditional banking system. Moreover, it was the product of thoughtful discussions around many initiatives that aim to diversify access to coding education for children around the world.

The focus of the discussion was on emphasizing decentralized AI governance to ensure equitable access, centering on the ownership of technology and data. It also covered innovations in AI teaching methods, notably through the use of decentralized teaching strategies employing affordable AI technology, to democratize education and knowledge dissemination in the field of artificial intelligence. Dr. Galfalvi continues…

“The population at the Foundation level is balanced in terms of gender inclusion and as a remote organization with no fixed location, there is a good global community of contractors and developers. But the next step is to include more people, everywhere.

This is where decentralization comes in. Everyone on the planet will be affected by AI, at least indirectly, and very soon (if not already). We’re working on decentralizing SingularityNET, but we are also focusing on creating ways for people to take part in discussions around global ethics and governance. That’s why we held a roundtable at the conference, where we discussed abstract notions of ethics and how they can be applied, but also practical systems that could be put in place to gather input from the global community.

We asked our participants to think of three questions that could stimulate meaningful discussion on AI ethics. Here are some of their contributions:

• What inputs are needed to make good decisions (taking into account giving a voice to that which is important but doesn’t yet have a voice)?
• What makes AI trustworthy?
• What are we optimizing for in human experience?
• If the definition of an ideal trusted AGI system includes the ability to opt out (among other things), although it is not possible now, could we design a system where people can opt out of being affected by AI in the future?
• If a truly inclusive system embraces the full diversity of ways of knowing, how might we provide for the full participation and validation of diversity of subjective values across ALL cultures?

These questions encapsulate the concerns and ethical paradoxes that persist in AI development. And this is just a sample; no one had trouble thinking of anything, which just goes to show how important these topics are for AI development in the future: these issues occupy everyone’s thinking. But they require more coverage and a bridge between thinkers (and that includes the general public) and AI developer communities.

We held these sessions partly because we are working on an initiative to include everyone’s views on ethics and AI governance. We are creating discussion spaces and will use AI tools to synthesize those discussions to provide insights into what the world thinks about AI and where it is going. By this time next year, we hope to have a way for people to have more agency in what AI is developed by providing ways to donate resources towards the projects they think are meaningful. We think that everyone should have a say in what is produced, not just the companies that are at the forefront of AI development.

Decentralizing governance and approaches to ethics provides more agency, and giving up sole custody of AI so that everyone can play a part in its development and governance.”

You can get involved too. Join the Mattermost group to stay tuned into the conversation on ethics, governance, and decentralizing beneficial AGI.