Generative AI tools like ChatGPT and Google's AI Overviews now process billions of queries daily despite widespread skepticism about their reliability. New data reveals only 8.5% of users fully trust AI-generated answers, yet adoption continues to surge as these tools reshape how we interact with information online.

The AI Trust Paradox: Why We Rely on Tools We Don't Believe In

Mininyx Doodle/Getty Images

Generative AI has become the internet's invisible infrastructure, processing 2.5 billion daily queries through ChatGPT alone—with 330 million originating from the US. This explosive growth positions OpenAI's chatbot to potentially challenge Google's 5 trillion annual queries, especially once its rumored browser launches. Yet beneath this adoption tsunami lies a troubling disconnect: users increasingly depend on tools they fundamentally distrust.

The Adoption Tsunami

ChatGPT became 2025's most downloaded app, surpassing TikTok, Facebook, Instagram, and X combined in June
Google responded with AI Overviews and AI Mode, embedding generative responses directly into search
Startup Perplexity entered the browser wars with Comet, challenging Chrome and Safari

"The picture that emerges is one where people interact with generative AI by default while placing little credence in its answers," observes Webb Wright, ZDNET contributing writer.

The Trust Gap

Recent surveys reveal a credibility crisis:

Only 8.5% of Americans "always trust" Google's AI Overviews

21% report zero trust in AI-generated answers

Over 40% rarely verify sources cited in AI responses

Trust fluctuates based on context—users distrust AI for legal/medical advice but may prefer it over humans for technical domains. Even response tone impacts credibility: sycophantic language reduces trust compared to neutral phrasing.

Why We Use Untrusted Tools

Three factors drive adoption despite skepticism:

Convenience override: Frictionless answers trump verification effort
Domain dependence: Willingness to trust varies by subject complexity
Behavioral inertia: AI becomes default interface through platform integration

The Path Forward

AI developers recognize the crisis. OpenAI and Anthropic now prioritize "explainable AI" initiatives to demystify outputs. Meanwhile, hallucinations remain endemic—the unsettling tendency for models to present fiction as fact.

As generative AI becomes the internet's new nervous system, the industry faces a critical challenge: building trustworthy systems for users who've already embraced imperfect tools. The future of information consumption hinges on closing the gap between utility and credibility.

Source: ZDNET - Webb Wright, July 2025