Trust but verify: The battle to fix the crisis of credibility

Something has shifted in how people relate to institutions. Across the OECD, more people now distrust their national government than trust it. In Canada, only 48% express confidence in the federal government, down from the high 50s before the pandemic.¹ An Ipsos survey captured the trajectory: trust in government to do what is right fell from 58% in 2019 to 43% by 2022.² Meanwhile, the 2025 CanTrust Index found that politicians are trusted by just 17% of Canadians, the lowest in a decade of tracking, and 6 in 10 say political parties are divisive forces.³

Social media and AI-generated content have accelerated the decline, with nearly half of Canadians now believing that AI will make information sources less trustworthy. Algorithms reward outrage over accuracy, flooding public discourse with polarizing content and AI-generated noise. As Jimmy Wales, the co-founder of Wikipedia, observed on a recent RBC Disruptors podcast, platforms incentivize bad behaviour through engagement: “you act like a jerk and you get engagement.”⁴

Wales’ latest book Seven Rules of Trust—A Blueprint for Building Things That Last, focuses on the global crisis of credibility and knowledge. Both are in short supply: The 2026 Edelman Trust Barometer found 73% of Canadians unwilling to trust someone with different values or information sources.⁵

The consequences of mistrust are far-reaching and having real impact: In Slovakia’s 2023 election, a deepfake audio clip impersonating a political party leader went viral during a legally mandated campaign silence period, leaving journalists no window to respond.⁶ In the United States, an AI-generated robocall mimicking President Joe Biden urged New Hampshire voters to stay home during the 2024 primary.⁷ Similar incidents surfaced in Bangladesh, Turkey, and India. The German Marshall Fund tracked 133 deepfake incidents tied to elections across dozens of countries.⁸

Wikipedia: An architectural proof of concept

Wikipedia makes for an instructive model. The free online encyclopedia covers more than seven million English-language articles, roughly 283,000 active editors, and billions of page views annually—all on a non-profit budget. It’s the go-to site for many to source everything from a storied company’s corporate history to oddities and obscure records.

For all its variety, it’s far from perfect: critics flag ideological biases, gender gaps among editors, and vulnerability to paid manipulation. But as Wales noted on the podcast, Wikipedia has gone “from being kind of a joke to one of the few things people trust.”

The reason is structural. Wikipedia’s model is “accountability, not gatekeeping,” Wales told RBC’s Disruptors podcast.⁹ “Everything you edit, everybody can see what you’ve done.” Every source is checkable, disputes happen on public talk pages, and corrections happen in real time.

Wales’s thinking was shaped early by Nobel-prizewinning philosopher Friedrich Hayek’s argument about decentralized knowledge—the idea that decision-making works best at the endpoints, not through a central hierarchy. Wales pointed to X’s Community Notes as a promising application of the same principle: empowering users rather than relying on top-down moderation.

Where this leaves Canada

Research going back to Knack and Keefer’s 1997 study confirms that trust is a measurable input to growth.¹⁰ A Deloitte analysis by chief global economist Ira Kalish makes the mechanism concrete: a rise in trust increases the quantity of business fixed investment, and it raises productivity through higher-quality investments, human capital accumulation, and greater internationalization.¹¹

The consultancy’s modelling suggests a ten-percentage-point increase in the share of trusting people within a country raises annual per capita GDP growth by about half a percentage point: a substantial gain when global growth averaged 2.2 percent between 2015 and 2019.

There is no single fix to restore trust in corporate and public sector governance. But as the Disruptors’ conversation with Wales highlighted, trust is not a moral decoration. The work of rebuilding it will be slow, uneven, and ongoing. But the cost of not starting is already measurable.