- Human.org has raised $7.3 million for safety tools to align AI with human values.
- The startup uses blockchain to verify identities and distinguish humans from AI agents.
- Backers in the funding round include HF0, Soma Capital, Spearhead, Pioneer Fund, and others.
A startup building a platform to ensure AI agents are aligned with human values and an app for people to verify they are human has raised $7.3 million.
San Francisco-based Human.org bills itself as an AI safety lab building “trust infrastructure” to keep AI accountable, said the startup’s 23-year-old founder and CEO, Kirill Avery.
“We’re two and a half years into AI-generated content, and it’s becoming impossible to differentiate which content has been created by humans and which is not — but this is just the beginning,” he told Business Insider.
With AI agents — software that can perform tasks autonomously —becoming more mainstream, Avery added that it’s vital that human expression is preserved and that AI remains aligned with human intentions.
AI alignment is the process of ensuring an AI’s actions are in step with the values of its human creators. The field gained mainstream traction as AI juggernauts raced to build high-performing models at breakneck speed, fuelling fears that the rise of a superhuman AI could disregard humanity’s key values.
Human is developing two core products. Its “Human Network” platform, powered by a blockchain protocol, is an infrastructure layer for interactions between humans and verified AI agents. The startup’s “Agent ID” protocol allows users to assign accountability to AI agents that can be traced back to their human creators. Because it’s on the blockchain, it isn’t controlled or owned by anyone and is accessible to the wider public.
Human’s other key product is the Human App, which Avery says is “basically the solution to proving that you’re a real human” and “helping people to show their humanity on different services.”
The startup has created two forms of online ID — one for humans and one for AI agents — that act as a “digital passport,” Avery told BI. “Signing in with Human ID is the essential part of knowing if this is a human or an agent you’re talking to, as a new form of ‘captcha’. The Human App is a place to manage your identity and all of your agent logs,” he added.
The $7.3 million pre-seed was backed by HF0, Soma Capital, Spearhead, Pioneer Fund, Hummingbird VC, Brett Gibson, general partner at Initialized Capital, and others.
With the fresh funding, Human will continue to focus on its deep tech research.
