Knowledge Graph Infographic

AI Has a PR Problem

The industry is winning the model race and losing the trust race. AI's net image is -20 — worse than ICE. The industry has accumulated institutional power without building institutional trust.

Author: Saanya Ojha
The Change Constant · May 5, 2026
-20AI net image — worse than ICE (-18). Only Democratic Party & Iran scored worse (NBC, Mar 2026)
26%Positive view of AI vs. 46% negative — despite 56% having used AI recently
57%Say AI risks outweigh benefits vs. 34% who say benefits outweigh risks
7Prescriptions for rebuilding trust — from language to labor to infrastructure
The Data

Utility ≠ Legitimacy

NBC News March 2026 poll of 1,000 registered voters: 56% used AI recently, but 57% say risks outweigh benefits. People are using AI but do not feel good about it.

Inside the Valley

Model capabilities, benchmarks, agents, cheaper inference, longer context windows. The conversation remains focused on technical progress — not public sentiment.

Outside the Valley

"Who asked for this?" — AI seen as imposed by a small group of companies with enormous capital, consuming energy, trained on others' work, threatening jobs, sold back as inevitable.

"The industry branded a technology that could help cure disease, tutor children, and defend networks in the same neighborhood as deportation raids and geopolitical enemies. That takes work."

— Saanya Ojha, May 2026

Backlash

Physical, Economic, and Personal

Economic Messaging

"You cannot tell people they are economically obsolete on Monday and expect them to applaud your product launch on Tuesday." AI entered the workplace wrapped in a threat.

Physical Backlash

Maine passed first-in-nation ban on data centers over 20 megawatts. AI became a land-use story — electricity, water, environmental impact, local control.

Personal Backlash

Suspect threw incendiary device toward Sam Altman's residence; OpenAI HQ threatened. AI leaders personalized as avatars of a feared future.

Core Diagnosis

From Inside: Abundance

Tools, efficiency, progress, innovation, productivity — AI looks like a net positive to those building and deploying it.

From Outside: Extraction

Scraped work, at-risk jobs, strained power grids, addicted kids, polluted feeds, vanishing trust — and the people doing it are becoming billionaires.

Solutions

Seven Prescriptions for a Better Social Contract

1. Stop Selling AI as Inevitability

The word is politically combustible — it tells people their consent is irrelevant. Replace with choice and democratic deliberation.

2. Be Honest About Tradeoffs

AI will create and eliminate jobs. The public needs candor about both, not fairy tales about universal abundance.

3. Make Benefits Visible and Local

Abstract GDP promises won't persuade a nurse, teacher, or small business owner who feels exposed. Deliver tangible local benefits.

4. Treat Labor Transition as Core

Not a philanthropic afterthought. Build real bridges from changed jobs to new ones — invest at infrastructure scale.

5. Take Infrastructure Politics Seriously

Communities deserve to know what they get for data centers, transmission lines, and water. Land-use politics are harder than app updates.

6. Build Trust Into the Product

Disclosure, provenance, privacy, auditability, and safety are adoption accelerants — not obstacles.

7. Use a More Human Vocabulary

Stop 'disrupting' people. Jobs aren't 'tasks.' Creative work isn't 'data.' Communities aren't 'sites.' Language shapes consent.

FAQ

Twelve Questions

Glossary

Key Concepts

AI Trust Crisis

Gap between technical progress and public legitimacy — net -20 image, worse than ICE.

Public Perception

56% use AI, 57% say risks outweigh benefits. The utility ≠ legitimacy paradox.

AI Social Contract

New compact: labor transition, infrastructure consent, honest tradeoff communication.

HowTo

1. Stop Selling Inevitability

Replace inevitability with democratic choice. The word tells people consent is irrelevant.

2. Be Honest About Tradeoffs

Acknowledge job creation AND elimination. Candor, not fairy tales.

3. Make Benefits Local

Concrete community benefits, not abstract GDP promises.

4. Labor Transition as Core

Real bridges at infrastructure scale, not philanthropic afterthoughts.

5. Infrastructure Politics

Communities deserve to know what they get for data centers, transmission lines, water.

6. Trust Into the Product

Disclosure, provenance, privacy, auditability — adoption accelerants.

7. Human Vocabulary

Stop 'disrupting' people. Language shapes whether people feel included or targeted.