51% | Will US public trust in online information decline by over 20 points in Edelman surveys from 2025 baseline by 2029?

Back to thesis map Part ofWill AI generated content exceed human generated content on the US internet before 2030?

Current eventSocial trust & institutional risk

Will US public trust in online information decline by over 20 points in Edelman surveys from 2025 baseline by 2029?

LiveResolves Dec 31, 2029

Current view

51%

Current 4 model consensus

Movement

Recent consensus movement

Model disagreement

38pt

Range across the model set

Evidence

Diverging

The evidence is sharply disputed

Thesis map

What is carrying or resisting the view?

Start with the map. Select a mechanism or condition to see what supports the thesis, what resists it, and where disagreement deserves review.

What changed

No decision-relevant findings are available yet.

Review record

Weekly freezes, daily history, and the map version preserved with the record.

Why the view is fragile

38pt model spread around 51% consensus. Open the full disagreement instrument only when you need the model-level reasoning.

Model disagreement

Model spread

38pt

32% to 70%

Consensus

51%

4 models

Evidence state

Diverging

Models sharply disagree

Most divergent

Gemini

+19pt from consensus

Disagreement synthesis

The models share a central case but assign different weight to the evidence.

Claude anchors the lower end at 32%, while Gemini anchors the upper end at 70%. Consensus sits at 51%, leaving a 38 point range for review.

Primary review question

The development and widespread adoption of effective AI detection and content authentication technologies by major platforms and governments could significantly mitigate the decline in trust.

Shared probability axis

Range 32% to 70%

Consensus 51%

Claude 32

Grok 37

GPT-4o 65

Gemini 70

0255075100

ModelForecastVs consensusDivergenceReasoning signal

Model reasoning

Select a model to inspect

Reasoning

Edelman's Trust Barometer has shown a general decline in trust in media and online information in the US in recent years, with significant year-over-year drops. Given the ongoing challenges of misinformation, the increasing sophistication of AI-generated fake content, and potential further political polarization, it's highly probable that trust will continue to erode, exceeding a 20-point decline from a 2025 baseline by 2029. For example, the 2023 Trust Barometer indicated a significant dip in trust for online news and social media, a trend that has persisted.

Key uncertainty

The development and widespread adoption of effective AI detection and content authentication technologies by major platforms and governments could significantly mitigate the decline in trust.

Deep dive

Full decomposition, component audit, trajectory, and detailed analytical views.

Deep dive

Detailed evidence, audit, and history views.

Open only what you need

Belief trajectoryConsensus 51% · 38pt spread

Belief trajectory0 updates

Belief trajectory

Consensus—

Spread—

Confidence—

Δ Consensus—

Loading belief trajectory…

Resolution & provenance

Resolution sources, record policy, cadence, and permanent record metadata.

Resolution and provenance

How the thesis resolves and how its record is maintained.

Permanent record

Record ID

CRENE-AICD-SO01

Record policy

Append only

Freeze cadence

Thursdays

Generated

Jul 8, 2026

Back to thesis map Part ofWill AI generated content exceed human generated content on the US internet before 2030?

Current eventSocial trust & institutional risk

Will US public trust in online information decline by over 20 points in Edelman surveys from 2025 baseline by 2029?

LiveResolves Dec 31, 2029

Current view

51%

Current 4 model consensus

Movement

Recent consensus movement

Model disagreement

38pt

Range across the model set

Evidence

Diverging

The evidence is sharply disputed

Thesis map

What is carrying or resisting the view?

Start with the map. Select a mechanism or condition to see what supports the thesis, what resists it, and where disagreement deserves review.

What changed

No decision-relevant findings are available yet.

Review record

Weekly freezes, daily history, and the map version preserved with the record.

Why the view is fragile

38pt model spread around 51% consensus. Open the full disagreement instrument only when you need the model-level reasoning.

Model disagreement

Model spread

38pt

32% to 70%

Consensus

51%

4 models

Evidence state

Diverging

Models sharply disagree

Most divergent

Gemini

+19pt from consensus

Disagreement synthesis

The models share a central case but assign different weight to the evidence.

Claude anchors the lower end at 32%, while Gemini anchors the upper end at 70%. Consensus sits at 51%, leaving a 38 point range for review.

Primary review question

The development and widespread adoption of effective AI detection and content authentication technologies by major platforms and governments could significantly mitigate the decline in trust.

Shared probability axis

Range 32% to 70%

Consensus 51%

Claude 32

Grok 37

GPT-4o 65

Gemini 70

0255075100

ModelForecastVs consensusDivergenceReasoning signal

Model reasoning

Select a model to inspect

Reasoning

Key uncertainty

The development and widespread adoption of effective AI detection and content authentication technologies by major platforms and governments could significantly mitigate the decline in trust.

Deep dive

Full decomposition, component audit, trajectory, and detailed analytical views.

Deep dive

Detailed evidence, audit, and history views.

Open only what you need

Belief trajectoryConsensus 51% · 38pt spread

Belief trajectory0 updates

Belief trajectory

Consensus—

Spread—

Confidence—

Δ Consensus—

Loading belief trajectory…

Resolution & provenance

Resolution sources, record policy, cadence, and permanent record metadata.

Resolution and provenance

How the thesis resolves and how its record is maintained.

Permanent record

Record ID

CRENE-AICD-SO01

Record policy

Append only

Freeze cadence

Thursdays

Generated

Jul 8, 2026