Meta, formerly known as Facebook, recently conducted a groundbreaking experiment in governance by involving over 6,000 people from 32 countries and 19 languages in a deliberative democratic process. This process aimed to inform decisions regarding Meta’s responsibility for the metaverse it is creating. The participants engaged in extensive conversations in small online group sessions and had the opportunity to hear from non-Meta experts on the relevant issues. Impressively, 82% of the participants expressed their willingness to recommend this format as a means for the company to make decisions in the future.
Building on the success of this experiment, Meta has publicly committed to implementing a similar process for generative AI, aligning itself with the growing interest in democratic innovation in governing or guiding AI systems. Other organizations, including Google, DeepMind, OpenAI, and Anthropic, are also exploring approaches based on deliberative democracy. This movement towards involving the public in decision-making aligns with the concepts promoted by proponents of democratic governance, such as myself.
As an advocate for this form of decision-making, I have had the opportunity to closely observe Meta’s Community Forums, the name given to these deliberative processes, since their establishment in 2019. Initially, I was involved in providing input on policies related to manipulated media. Around the same time, I became aware of representative deliberations, an approach in which governments involve a representative group of citizens to make decisions on complex policy questions. This method provided a potential solution for making decisions about technologies that impact people across national boundaries. I began urging companies to pilot such processes to address their most challenging issues. Witnessing Meta independently initiate this pilot was truly exciting.
Meta’s mammoth 32-country Community Forum was noteworthy because it demonstrated that running this kind of process is feasible, despite the logistical challenges. The proceedings were largely managed by Meta’s partners at Stanford, and there was no evidence of any attempt by Meta employees to influence the outcome. Furthermore, Meta stayed true to its commitment to have the Stanford partners directly report the results, regardless of their nature. It was evident that some individuals within the company were genuinely evaluating how best to implement the potential outputs of the forum. These outputs included perspectives on appropriate repercussions for instances of bullying and harassment in Metaverse spaces and suggestions for moderation and monitoring systems.
Moving forward, there are areas where improvement is needed to ensure a truly democratic process. Participants should have greater power and agency, and the process itself should be more transparent and publicly accessible. These enhancements will enhance the democratic nature of the decision-making processes and provide participants with a stronger voice.
The success of Meta’s venture into democratic governance through its Community Forums is an encouraging proof of concept for transnational democratic governance. As Meta turns its focus to generative AI, it is poised to foster greater public involvement in shaping decisions that impact society at large. This step aligns with the growing trend of involving citizens in the governance of AI systems. As an active participant in this movement, I look forward to further advancements and refinements in this space, facilitating a more democratic and inclusive future.
