Public trust in artificial intelligence rose from 26% in 2025 to 30% in 2026, according to CARMA’s latest study, ‘Perceptions of AI: A Power Shift,’ which examines media coverage and public opinion of artificial intelligence across the first half of 2026.
The research is the second installment in a series by CARMA tracking global AI sentiment. This edition analyzed over 6,000 articles across 500 tier-one media outlets in 45 markets, and integrated results from a 6,300-person public opinion survey spanning 19 countries.
Trust overtook excitement as the public’s dominant emotion towards AI for the first time, registering at 33% against 29%. Media trust signals fell from 35% to 26% over the same period.
Governments and the public disagree on responsibility when AI gets it wrong
Media coverage placed the greatest share of responsibility, 35%, for AI’s future on governments and regulators. The public placed governments last, at 10%, and put AI companies and developers first, at 33%.
CEOs continued to be the strongest advocates of AI, but academics were the most trusted voices, suggesting audiences value expertise and independence over optimism when forming opinions about AI.
CARMA cites a disagreement that played out between two prominent figures in the same week. Argentina’s President Javier Milei wrote in the Financial Times that AI required regulatory freedom to reach its economic potential. Four days later, historian Yuval Noah Harari wrote in the same publication, warning that granting AI systems legal rights without human accountability could let them operate beyond anyone’s control.
According to CARMA, the findings indicate that public expectations of accountability are shifting toward industry, independent of regulatory timelines.
AI buzzwords are losing ground to numbers
59% of respondents believe companies exaggerate their AI use. 44% say they pay less attention to any message that mentions AI at all.
The public prefers simpler messaging: understanding how the technology works, and how to measure productivity, efficiency, and accuracy.
CARMA’s findings show that media coverage has tended to link trust in AI to safety and reliability. The public, however, placed greater weight on transparency and on understanding how AI systems reach their decisions.
Concerns Remain Despite Rising Trust
Cybercrime, misinformation, and job displacement emerged as three leading concerns, with CARMA noting that, unlike most other findings in the study, media narratives and public opinion converged closely on these issues rather than diverging.
A recent Pew Research Center survey highlights why rising trust in AI does not mean concerns have disappeared. Among adults under 30 in the US, 55% now say they are more concerned than excited about AI, reflecting continued uncertainty around the technology’s broader impact.
Pew also found that concerns around AI have increased across age groups, with excitement declining since 2021. The survey found that 71% of US adults believe AI will lead to fewer jobs over the next two decades, rising to 73% among adults under 30.
Anthropic share of voice rises
According to the CARMA report, Anthropic’s long-term positioning around AI safety and responsible development gained traction following its high-profile clashes with the US government. The company’s share of voice rose from 4% to 17%, overtaking Google to become the second most visible AI company in global coverage.
