Public trust in AI rises as media becomes more sceptical, CARMA finds

Public trust in artificial intelligence is rising even as media coverage becomes more sceptical, according to new research from media intelligence firm CARMA.

Its latest study, "Perceptions of AI: A power shift from productivity tool to strategic infrastructure", found that public trust in AI rose from 26% in 2025 to 30% in 2026. At the same time, media trust signals fell from 35% to 26%, highlighting a growing gap between how the technology is portrayed in the press and how audiences perceive it.

The research analysed 6,006 articles from 500 tier-one media outlets across 45 markets between January and June 2026, alongside a public opinion survey of 6,300 people across 19 countries, including Indonesia, India, China and Australia.

Don't miss: Creators cracked the algorithm. Can they crack AI search?

For the first time in CARMA's tracking, trust also overtook excitement as the public's leading emotional response to AI, at 33% versus 29%. This suggests audiences are moving beyond the novelty of AI and becoming more comfortable with its practical applications.

The shift comes as media coverage itself enters a more critical phase. While positive narratives still accounted for the majority of AI coverage, optimism fell from 57% in 2025 to 47% in the first half of 2026. Negative coverage increased from 13% to 28%.

Rather than focusing primarily on new model launches and technological breakthroughs, coverage increasingly examined how AI should be deployed, who should control it and what its implications are for governance, national security and society.

Trust is moving beyond the media

CARMA's findings suggest that people's own experiences with AI may increasingly be influencing perceptions alongside traditional media narratives.

The research found that audiences were more likely to associate trust with transparency around how AI works and how decisions are made, while media coverage placed greater emphasis on safety, reliability and accountability.

This distinction could have implications for brands and organisations communicating their AI strategies. Demonstrating that an AI system is safe may no longer be enough; audiences also want to understand what happens behind the technology.

The demand for greater transparency comes amid widespread scepticism towards corporate AI messaging. Some 59% of respondents said companies exaggerate their use of AI and rely on buzzwords, while 44% said they pay less attention to messages that mention AI.

By contrast, messages focused on productivity, efficiency, accuracy, speed and scalability were viewed as more meaningful.

Accountability gap emerges

One of the clearest differences between media coverage and public opinion concerned who should be responsible when AI systems cause harm or make mistakes.

Media coverage assigned the largest share of responsibility to governments and regulators, at 35%. The public took a different view, placing AI companies and developers first at 33%, while governments and regulators ranked last at 10%.

Both groups nevertheless identified AI companies and developers as important actors in determining the technology's future.

The divide comes as AI governance becomes increasingly tied to geopolitics and national competitiveness. CARMA found that coverage had shifted towards issues including technological sovereignty, export controls, semiconductor competition and national AI infrastructure.

AI's role in defence also became more prominent during the first half of the year. However, the research found that coverage was focused less on autonomous weapons than on AI's broader strategic role. National security and strategic competition together accounted for more than two-fifths of coverage concerning AI's role in defence.

Anthropic's visibility surge highlights the role of action

The changing AI landscape was also reflected in the companies receiving media attention.

OpenAI remained the most visible AI company, accounting for 21% of share of voice. Anthropic, however, recorded the biggest increase, climbing from 4% in 2025 to 17% in the first half of 2026 and overtaking Google to become the second most visible AI company in global coverage.

CARMA attributed the rise not to a communications campaign, but to Anthropic's public clashes with the US government over military AI and export controls.

The episode brought greater attention to the company's long-standing positioning around AI safety and responsible development, illustrating the gap between communicating a corporate position and demonstrating it through action.

The report also found that CEOs remained the most prominent advocates of AI in media, generating the most positive coverage. Yet academics were the most trusted voices among audiences, suggesting that expertise and perceived independence may carry more weight than optimism alone.

Cybercrime, misinformation and jobs dominate concerns

Despite the rise in public trust, significant concerns around AI remain.

Cybercrime, misinformation and job displacement emerged as the three leading drivers of AI-related distrust among audiences. These were also among the most prominent concerns in media coverage, marking an area where media narratives and public opinion were closely aligned.

Job displacement was particularly significant for younger audiences. CARMA found that coverage of AI's impact on younger generations was overwhelmingly negative, with stories focusing on employment, education, authenticity and mental wellbeing.

The research also found that AI was increasingly associated with workforce transformation in coverage of CEOs. Eleven per cent of CEO coverage about AI appeared in the context of layoffs, with more than half of those articles carrying a negative tone.

As public trust continues to rise, CARMA's research indicates that audiences are not necessarily becoming less critical of AI. Instead, they appear to be becoming more discerning about how it is used, who is responsible for it and whether organisations can demonstrate its value.

Discover what's next in Indonesia's fast-moving digital landscape at DMA Indonesia 2026 on 13 October in Jakarta, where marketing leaders will explore how AI, creator ecosystems, social commerce, and omnichannel strategies are reshaping customer engagement and business growth.

Related articles:
AI watermarking is raising the value of human writing, not killing it
Meta AI gets new tools to analyse and optimise ad campaigns
Is your brand caught in the AI slop? Consumers are taking notice