Trust in scientists holds. The challenge is how science reaches the public.

Science has never moved faster. Artificial intelligence is accelerating discovery, transforming research and changing how scientific knowledge reaches the public. Yet as science advances, the environment in which it is communicated is becoming increasingly complex.

The familiar narrative of declining trust in science tells only part of the story. Research involving more than 71,000 people across 68 countries shows that trust in scientists remains moderately high. However, confidence in the information people receive about science is more fragile. In the UK for example, only 40% of people feel confident that what they hear about science is generally true.

AI is adding a new dimension to this challenge. As people increasingly turn to AI assistants for scientific information, questions around accuracy, credibility and the attribution of research are becoming more pressing.

Our latest report, “Trusting scientist, mistrusting science: Communicating discovery in the age of AI”, prepared in the context of the Geneva Science and Diplomacy Anticipation Summit, examines what these developments mean for scientific organisations, universities and NGOs.

Combining international research with Leidar’s analysis of media and social media conversations, the report identifies three significant trends over the past twelve months.

• Mentions of AI in scientific discovery increased by 90%.
• Discussions of AI as a source of health and science advice rose by 190%.
• Mentions of AI-fabricated citations and studies surged by 506%.

These findings point to a growing challenge for organisations producing scientific knowledge. As discoveries travel through increasingly fragmented information channels, credibility depends not only on the quality of the science, but also on how it is explained, who communicates it and whether its original source remains visible.

For scientific organisations, this calls for a more strategic approach to communications. Building trust means being transparent about uncertainty, ensuring researchers remain credible public voices and understanding how scientific findings are interpreted, challenged and shared, including through AI.

The report outlines four practical steps to help organisations protect their credibility, strengthen their visibility and prepare for greater scrutiny.

As AI continues to reshape both scientific discovery and public understanding, organisations that invest in how their knowledge is communicated and attributed will be better positioned to maintain trust.

LN – Trusting scientist mistrusting science report