Self Labs has introduced a new workspace, Sovereign, to allow companies, governments, municipalities, and organizations implement verified-human evals, access gates and distribution mechanisms for Sovereign AI more easily requesting users’ proof of citizenship, in a fair and privacy-preserving manner.

When export controls pulled Fable 5 and Mythos 5 offline days after launch, access to some of the most capable models on earth was suddenly sorted by passport. In that moment, sovereign AI went from a talking point to a global security question. Companies looked to address the need for identity and residency verification, and governments throughout the world examined what responsible, fair access to AI would look like in the real world. The need was clear for strong solutions that protect users while serving the regulatory and compliance requirements of the organizations deploying it.

Self’s infrastructure for the agentic web was uniquely built for this. The new Sovereign Workspace helps organizations put those tools to use, enforcing their access policies without ever collecting or storing user data.

Self’s Sovereign Workspace enables several AI use cases, including: 

• Access by nationality or residency. Enforce citizens-first or residents-first policies for model access without collecting a single passport scan. Users do not need to reveal their personal data, only the proof that attests to their eligibility. 

• Fair allocation of compute. Proof of human is the foundation for distributing compute per person rather than per account, the same primitive behind Universal Basic Compute. One person cannot claim ten allocations, and no one has to hand over their identity to claim one.

• Verified-human pipelines. Sovereign models still need humans to train and evaluate them. The data that teaches models what good looks like is increasingly polluted by synthetic identities. Attaching verified-human attestation to labeling, RLHF, and red-teaming work makes the pipeline trustworthy from the start.

• Human and agent distinction. Deployments can tell which requests come from a person, and every agent stays accountable to the human behind it.

Self Enterprise gives organizations a direct line to Self's team and engineering resources, and to leverage the infrastructure already trusted by Google, Opera, and 18M users across 180+ countries. Self is working with various partners on sovereign AI today, with more to share as those projects mature.

Try the Sovereign workspace today for free with Self Enterprise at dashboard.self.xyz

For any questions, reach out to team@self.xyz