Why we invested in AMMO, a community-driven AI ecosystem

In an era of expanding digital content, users and developers face a fundamental dilemma: centralized platforms control what gets created, how it’s moderated, and how users interact with it. These opaque moderation policies and profit-driven recommendation algorithms often erode trust, limit meaningful customization, and stifle new forms of collaboration. AMMO aims to solve this issue by building a community-driven AI ecosystem that empowers content creators and consumers to shape, moderate, and evolve the media they engage with.

AMMO is an early-stage AI platform that envisions a network of specialized multi-agent systems working in unison to generate, govern, and iterate on content. This design leverages human feedback, agentic frameworks, dynamic policy, and reinforcement learning (RL) reward functions to continuously refine both individual AI agents and the system as a whole. Incentive mechanisms, such as tokens, further encourage collaboration and alignment across these AI workers.

Within AMMO’s multi-agent architecture, discrete tasks like scripting, editing, quality checking, and directing are delegated to specialized AI “workers.” Meanwhile, governance agents uphold policies, community standards, and ethical guidelines. This self-improving setup aspires to serve user interests first rather than chasing engagement-driven metrics.

We believe AMMO’s potential is evident in its founding team’s strong AI credentials. David Huang (Co-founder and CEO) spent a decade at Google, including seven years leading AI initiatives and strategic services in mobile. Diego Hong (Co-founder and CTO) previously led consumer chatbot efforts at Meta and earned his degree from Oxford. Two other co-founders each hold PhDs with significant expertise in multi-agent systems and GPU computing. Collectively, they bring numerous NeurIPS publications, AI patents, and extensive machine learning research, forming a powerful combination of technical rigor and consumer-oriented vision.

At its core, AMMO’s architecture merges cutting-edge AI techniques in content summarization and moderation with robust, zero-trust, community-led governance. In the near term, AMMO’s prototype will enable creators and everyday users to produce and fine-tune content via multiple AI agents—each specializing in tasks like editing or scriptwriting—while policy agents enforce guidelines. Over time, these agents learn from direct user feedback and performance data, creating a virtuous cycle of continuous improvement and alignment. By employing Reinforcement Learning from Human Feedback (RLHF), AMMO refines not only individual outputs but also the entire network of AI agents and their interactions.

Ultimately, AMMO confronts the challenge of centralized, opaque moderation policies that can undermine user trust. By championing a multi-agent, community-centric model, AMMO offers a transparent alternative in which governance is shared and AI tools enhance, rather than overshadow, the user. We’re excited to back the team and believe in AMMO’s potential to shape this emerging AI ecosystem and integrate advanced AI capabilities across a diverse range of devices and consumer experiences.

Sam Campbell is an investor at Samsung Next. Samsung Next’s investment strategy is limited to its own views and does not reflect the vision or strategy of any other Samsung business unit, including, but not limited to, Samsung Electronics.