The Cosmos Institute, a nonprofit whose founding fellows include Anthropic co-founder Jack Clark and former Defense Department technologist Brendan McCord, has announced a venture program and research initiatives to — in the organization’s words — “cultivate a new generation of technologists and entrepreneurs equipped with deep philosophical thinking to navigate the uncharted territory of our AI age.”
In a blog post, McCord, Cosmos’ chair, said that the Institute will found an AI lab at the University of Oxford called the Human-Centered AI Lab, or “HAI Lab” for short. It’ll be led by Oxford philosopher Philipp Koralus — another of the Institute’s founding fellows — and will aim to “translate the philosophical principles of human flourishing into open source software and AI systems.”
What does “translating the philosophical principles of human flourishing” entail, exactly? Unclear. But the crux is to foster the creation of AI tech that respects human dignity while avoiding promoting harmful disruption (e.g., automation leading to job loss). A high and nebulous bar, to be sure — but one the HAI Lab is going to strive for.
The Institute is also spinning up a fellowship — the Cosmos Fellowship — with a cohort of four fellows to start. Working at the HAI Lab or other partner institutions for between “a term and a year,” fellows will collaborate with Cosmos mentors and pursue independent projects to “[explore the] intersection of AI expertise and deep philosophical insight,” McCord said.
There’s an investment component to the Institute’s plans, and it’s launching in the form of a venture org: Cosmos Ventures. Led by former DeepMind product lead Jason Zhao, ex-Stripe head of corporate strategy Alex Komoroske, Darren Zhu, and Zoe Weinberg, Cosmos Ventures will support “provocative new prototypes, essays and creative projects that explore fundamental questions around the philosophy of technology,” McCord said.
McCord characterizes Cosmos Ventures as “low overhead,” modeled after Institute fellow Tyler Cowen’s Emergent Ventures. Investments will range between $1,000 and $10,000 per project, and projects — which can be new or existing — must produce a “major deliverable” within three months.
The first group has already been funded, McCord said.
The Cosmos Institute isn’t the first to attempt to advance a more ethical, humanistic vision of AI. OpenAI, founded with the mission of delivering the benefits of advanced AI to all humanity, has dismantled entire safety teams. Clark’s own Anthropic once positioned itself as a safer, more ethical vendor, but in recent months has pushed back on AI regulations and aggressively scraped data without permission.
Perhaps the Cosmos Institute will fare better with its embrace of what McCord calls “accelerationism”: the idea that the future isn’t predetermined and therefore humanity is responsible for what it holds.
“We must build AI that encourages inquiry over complacency and promotes active engagement over passive dependence, especially in education, medicine, the public square and other vital domains,” McCord writes. “We champion AI that decentralizes power and enables bottom-up solutions, allowing individuals and communities to co-create a richer, more diverse society.”
They’re admirable positions. This reporter only hopes that money, influence and power don’t make the temptation to abandon them irresistible.
