How is trust affected by AI-generated content?

Trust in AI-generated content is a rapidly evolving issue, and it's deeply entangled with questions of credibility, transparency, and control. Here's a structured look at how trust is being affected:

1. The AI trust gap

Widespread scepticism: A recent survey found that 82% of users are sceptical of AI-generated content, yet only 8% consistently fact-check it.
Inconsistent quality: Over 42% of users have encountered inaccurate or misleading information in AI-generated summaries, especially in tools like Google’s AI Overviews.
Engagement paradox: Despite low trust, many users still rely on AI content for convenience – highlighting a tension between efficiency and credibility.

2. Cognitive vs. Affective Trust

A study in The Electronic Library breaks trust into two dimensions:

Cognitive trust: Based on perceived intelligence, transparency, and accuracy. This is undermined by knowledge hallucinations and algorithmic bias.
Affective trust: Influenced by perceived empathy and user experience. If users feel the AI “understands” them, they’re more likely to trust it - even if it’s wrong.

3. Transparency isn’t enough

Disclosure fatigue: Labelling content as AI-generated doesn’t always help. Mozilla’s research found that human-facing labels (like “written by AI”) are often ignored or misunderstood.
Machine-readable watermarks are more secure but require robust detection tools to be effective. Without them, even well-intentioned transparency efforts fall short.

4. The rise of “AI slop”

The internet is being flooded with low-value, AI-generated content – sometimes called “AI slop.” This includes fake reviews, shallow articles, and spammy SEO content.
This glut is eroding trust in search engines, recommendation systems, and even journalism, as users struggle to distinguish signal from noise.

5. What builds trust in AI content?

Trust Factor

Positive Impact

Negative Impact

Transparency

Clear attribution, visible disclosures

Hidden or misleading authorship

Accuracy

Fact-checked, human-reviewed content

Robotic or generic tone

Empathy

Personalised, context-aware responses

Robotic or generic tone

Consistency

Reliable performance across topicsReliable performance across topics

Erratic or biased outputs

Verification Tools

Watermarks, source links, citations

Lack of provenance or traceability

Where This Is Headed

Trust will become a differentiator: Platforms, such as REDOOR, that can show AI content is accurate, transparent, and responsibly sourced will gain user loyalty.
Hybrid models may win: Combining AI efficiency with human editorial oversight could offer the best of both worlds.
Regulation is coming: Expect more pressure for watermarking, disclosure standards, and AI content governance.

This resource was partially assisted by Microsoft Copilot.