AmericaDo Americans want oversight of AI that makes important decisions?
This survey found majority support across parties for testing new high-impact AI systems. The related H.R. 5511 Algorithmic Accountability Act of 2025 addresses impact assessments, not an identical government-approval test. Its official record checked October 9, 2026 showed committee referral.
PPC national Q73
Government testing of new high-impact AI systems.
U.S. adults; weighted opt-in panels; total n=2,509; four-fifths modules n=1,991–2,020 (question-specific n unreported). · 2026-02-11 / 2026-02-26. These are survey results, not votes on the linked bill. Missing breakdowns are not zero.
Record essentials
- Mechanism in the introduced bill
- FTC-required impact assessments of certain algorithms Verify this field ↗
Assessment does not automatically mean approval
A developer’s assessment, an independent audit, a regulator’s test and a mandatory authorization are different mechanisms. A meaningful comparison states who evaluates the system, when it happens, what is measured, and what follows a failure.
What this related bill does and does not prove
H.R. 5511 would direct the FTC to require specified impact assessments. That is a related accountability mechanism, not proof that the poll’s exact test has been implemented. Read covered-entity and system definitions before applying the proposal to a particular product.
Questions an affected person should be able to ask
Who owns the decision? What evidence can be corrected? Is a human review available? What record documents a challenge? These questions identify a practical accountability pathway without assuming every institution offers the same remedy today.
What you can do next
- Identify the real decision maker and affected process.
- Compare independent testing, internal assessment and appeals separately.
- Check the dated proposal and relevant institution’s current procedures.
This page does not submit a comment, register you to vote, allocate money, or contact an official on your behalf.
Keep this boundary in view
Support for a described test is not evidence of technical efficacy or support for this bill as a whole. This page does not state a universal right to appeal every automated decision.
A selected pilot, not an exhaustive policy platform. Survey findings describe the sampled population and the stated question, not all Americans or every version of a proposal.
Primary sources
- PPC AI polling release ↗2026-10-09 · Consulted October 9, 2026
- PPC questionnaire, Q73 ↗2026-10-09 · Consulted October 9, 2026
- Official H.R. 5511 action record ↗2026-10-09 · Consulted October 9, 2026
- H.R. 5511 introduced text ↗2026-10-09 · Consulted October 9, 2026
Questions worth resolving
Is an impact assessment the same as government approval?
No. Who performs the assessment and what happens afterward determine its practical meaning.
Does the poll measure whether a particular AI test works?
No. It measures support for the stated proposal, not system performance.
Keep your own evidence notes.
Use this space to record a source, an unresolved question, or a next step. Nothing is uploaded or saved by America.co. Avoid entering sensitive personal information.