Brad Smith on Require human-in-the-loop oversight for…

Brad Smith

Microsoft vice chair and president

We have to ensure that A.I. remains subject to human control. Whether it’s a government, the military, or any kind of organization, that is thinking about using A.I. to automate, say, critical infrastructure, we need to ensure that we have humans in control, that we can slow things down or turn things off.

AI Verified source (Aug 29, 2023)

4mo ago

Policy proposals and claims

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

votes For

Statement relation verification history AI Verified Report this

AI Verified The quote clearly supports the full statement: it says AI in high-stakes settings like government, the military, and critical infrastructure must remain under human control, with humans able to slow or shut it off. That implies support for requiring human-in-the-loop oversight in high-stakes domains. · YouCongress gpt-5.4-2026-03-05 · 4mo ago

Vote inference verification history AI Verified Report this

AI Verified The quote clearly supports requiring human oversight in high-stakes uses: it says A.I. must remain "subject to human control" and that for uses like "the military" or "critical infrastructure," "we need to ensure that we have humans in control." · YouCongress gpt-5.4-2026-03-05 · 4mo ago

Quote authenticity verification history

Report this

AI Verified The Fortune article at the provided URL contains this quote verbatim and explicitly attributes it to Microsoft president Brad Smith, saying he "told CNBC’s Martin Soong." The article is dated August 29, 2023, so the stored author, date, source URL, and quote text all match reliable evidence. · YouCongress gpt-5.4-2026-03-05 · 4mo ago

Disputed Brad Smith did say this in substance, but the supplied citation is wrong and the wording is not fully verbatim. The cited The National page from February 26, 2026 is about “collective defence” and does not contain this passage; searches on that page find no matches for “human control,” “critical infrastructure,” or “turn things off.” Reliable 2023 reports in Fortune (August 29, 2023) and IBTimes UK (September 1, 2023) attribute the line to Smith’s CNBC interview, but their wording includes extra words omitted here, such as “that is” and “say,” before “critical infrastructure.” ([thenationalnews.com](https://www.thenationalnews.com/future/technology/2026/02/26/microsoft-brad-smith-ai-on-the-record/)) · YouCongress gpt-5.4-2026-03-05 · 4mo ago

replying to Brad Smith

Other opinions from Brad Smith

See all

Brad Smith votes For

Require AI labs to publish safety evaluations before deploying frontier models

Frontier models do need to be assessed before release.

AI Verified source (Jul 9, 2026)

added 2mo ago

Brad Smith votes For

AI value alignment must be subject to public deliberation, not decided by developers alone

We do need an approach that is inclusive, that makes room for every government in the world, as well as multi-stakeholder conversations as were happening right now today. Government cannot act alone. The private sector cannot act alone.

AI Verified source (Jul 7, 2026)

added 3mo ago

Brad Smith votes Against

AI will create a permanent underclass

There was a time when humanity discovered that a horse could run faster than a person. So, people learned how to ride horses. Let's use AI to help people do more and not replace us.

AI Verified source (Jun 16, 2026)

added 2mo ago