Thomas Larsen on Require frontier AI labs to conduct nea…

Thomas Larsen

AI Futures Project researcher writing on Plan A assumptions, AI governance, and AI safety strategy.

Proposal #1: Total Research Transparency. Nearly all AI research is fully transparent to the public, including AI algorithms, code, and documentation, for all frontier AI projects. AI model weights, significant fractions of the training data, and a small amount of other sensitive information is prevented from leaving the datacenters. [...] Overall, we think that poor regulation is more concerning than covert projects, and so we recommend Total Research Transparency.

AI Verified (Jul 13, 2026)

Source:

Thomas Larsen, “Transparency Plan,” AI 2040 / AI Futures Project, published July 13, 2026. The single-author policy supplement presents and recommends “Total Research Transparency”: nearly all frontier-AI research is public while model weights and other sensitive information are prevented from leaving datacenters.

1h ago

Policy proposals and claims

Require frontier AI labs to conduct nearly all R&D publicly but keep the weights closed

votes For

Statement relation verification history AI Verified Report this

AI Verified The quote expressly covers nearly all frontier-AI R&D being public while model weights and sensitive information are kept in datacenters. · Hector Perez Arenas gpt-5 · 54min ago

Vote inference verification history AI Verified Report this

AI Verified For: Larsen recommends a regime with nearly all frontier R&D public while weights remain confined to datacenters. · Hector Perez Arenas gpt-5 · 54min ago

Quote authenticity verification history

Report this

AI Verified The July 13 AI 2040 Transparency Plan is a single-author Thomas Larsen page and contains the stored Proposal #1 wording and recommendation. · Hector Perez Arenas gpt-5 · 54min ago

replying to Thomas Larsen

Other opinions from Thomas Larsen

See all

Thomas Larsen votes For

We should adopt AI 2040 Plan A

We recommend scaling quickly to AIs that are as capable as top human experts, and only then pause as the world is being widely transformed.

AI Verified source (Jul 11, 2026)

added 23d ago

Thomas Larsen votes For

The U.S. and China should establish Mutually Assured Compute Destruction

Finally, I’m also excited about scenarios for plans that are less desirable than Plan A which may be easier to implement.

AI Verified source (Jul 11, 2026)

added 2mo ago

Thomas Larsen abstains

Pausing AI at human-genius level to solve alignment is safer than either racing to superintelligence or halting entirely

I’m confident that a cooperative international approach (Plan A) is better than the alternatives because the alternatives involve being in a race during the intelligence explosion, which incurs huge risks. I’m not confident about the exact variant of...

more AI Verified source (Jul 9, 2026)

added 21d ago