02 / Common Ground

Shared aims. Different routes.

Who gets to check that AI is safe? A close reading of licensing and openness, and the questions that sit between them.

Explore this pair in the Atlas

A / 25 Oct 2023

AI Licensing for a Better Future

Encode Justice & Future of Life Institute

Inspect the document

B / 31 Oct 2023

AI Safety and Openness

Mozilla & a group of signatories

Inspect the document

Editorial comparison of dated texts. These are not mutually exclusive camps. Shared words do not establish shared ideology or identical policy preferences.

Supported overlap / editorial reading

Make powerful systems accountable.

Both documents want to reduce AI harms and subject powerful systems to scrutiny. That gives us a concrete starting question: who needs access to what evidence, and who can act on what they find?

The boundary

Agreement has a scope.

Some licensing and openness arrangements could work together. Others could conflict. These texts do not establish agreement about the exact release conditions, authority or distribution of costs.

Make it concrete.

Six questions beyond the slogan

Who is meant to benefit?

A / AI Licensing for a Better Future

The licensing letter addresses harms to people and communities, including present harms and future risks.

B / AI Safety and Openness

The openness statement emphasizes public accountability, research access and responsible new entrants.

Where that leaves us: Partial overlap. Neither document supplies a complete beneficiary-by-beneficiary impact assessment.

Explore the mechanism

Design the conditions
for accountability.

A hypothetical lab about authority, access and appeal. Explore consequences without attributing consent to real groups.

Enter the lab

Another comparison

Same concern. Same proposal?

Concern about AI risk does not automatically imply support for a particular pause.

Read the comparison

Prepared 25 September 2026 · Sources: document A and document B. Private pilot interpretation; see editorial methods.