Brad Carson on AI labs should conduct risk assessments…

Brad Carson

Americans for Responsible Innovation president

ARI recommends targeting developers as a proxy for regulating frontier models. This is preferable to a regime targeting individual models, since today’s frontier systems derive their capabilities from a combination of baseline models, scaffolding, and compute; therefore any threshold defined by model characteristics alone may become obsolete as the field rapidly advances. [...] Additional risk domains such as autonomous harmful conduct; oversight subversion and deceptive alignment; autonomous self-replication and resource acquisition; autonomous AI R&D and high impact internal deployment; and manipulation and influence operations should be considered as additional areas of interest and potential review before deployment is approved.

AI Verified (May 11, 2026)

Source:

Brad Carson, “Letter to White House: Guidance on Oversight of AI Models,” Americans for Responsible Innovation, May 11, 2026, pp. 1–2.

20d ago

Policy proposals and claims

AI labs should conduct risk assessments before high-stakes internal deployment of frontier models

votes For

Statement relation verification history AI Verified Report this

AI Verified The quote identifies frontier systems/developers and calls for review of autonomous AI R&D and high-impact internal deployment before approval, supporting pre-deployment risk assessment. · Hector Perez Arenas gpt-5 · 20d ago

Vote inference verification history AI Verified Report this

AI Verified The quote supports pre-approval review of high-impact internal deployment and frontier-system risk domains; recorded answer for matches. · Hector Perez Arenas gpt-5 · 20d ago

Quote authenticity verification history

Report this

AI Verified ARI’s 2026-05-11 letter reproduces this wording on pp. 1–2; it is authored by Brad Carson and includes the cited pre-deployment-review language. · Hector Perez Arenas gpt-5 · 20d ago

replying to Brad Carson

Other opinions from Brad Carson

See all

Brad Carson votes For

Mandate third-party audits for major AI systems

By establishing robust safety standards for developers, requiring independent assurance of those standards, and ensuring transparency into frontier AI development, the federal government can give the public meaningful protections while preserving Ame...

more AI Verified source (Aug 10, 2026)

added 2mo ago

Brad Carson votes For

Require large datacenters to install kill switches for AI containment

Advanced AI models should never be deployed without a reliable off switch. The AI Kill Switch Act establishes a commonsense safeguard by requiring leading AI companies to maintain the ability to shut down their models and empowering the federal gover...

more AI Verified source (Jul 23, 2026)

added 3mo ago

Brad Carson votes For

Require human-in-the-loop oversight for agentic AI systems acting in high-stakes domains

Advanced AI models should never be deployed without a reliable off switch. The AI Kill Switch Act establishes a commonsense safeguard by requiring leading AI companies to maintain the ability to shut down their models and empowering the federal gover...

more AI Verified source (Jul 23, 2026)

added 3mo ago