CARMA, the global media intelligence and marketing/public relations communications analysis firm, recently published an eye-opening study comparing media messaging about artificial intelligence (AI) with public perceptions about AI, and what it found was that there is a significant gap between the two. Comparing 2025 to the first half of 2026, CARMA found that the media narrative has grown more negative, but at the same time, public enthusiasm and trust in AI has increased. The findings, surprising as they are, have important implications for policymakers, businesses, the media and the general public.
To understand the context of the CARMA study, it helps to understand what the company does. In a nutshell, CARMA’s business is to, in its own words, “analyze what the media message is, and research the market to discover who believes it,” so that its clientele can improve their marketing and public relations communications.
From the media analysis side, the study found that the broader narrative has shifted in the first half of this year compared to 2025, focusing less on the technological innovation of AI and its potential effects on productivity and business transformation, and more on issues of trust, governance and geopolitical competition. “Compared with 2025, coverage became less focused on what AI could do and more concerned with how it should be deployed, who should control it, and what role it would play in national competitiveness and security,” the study said.
The study assessed media coverage according to whether it was “positive,” “negative,” “balanced” or “factual.” On that scale, there were big changes from 2025 to the first half of this year. Positive coverage decreased from 57 percent of all stories to 47 percent, while negative coverage increased from 13 percent to 28 percent. Balanced coverage presenting both the positive and negative aspects of stories dropped from 26 percent to 12 percent, while strictly factual reporting increased from 4 percent to 12 percent. These two latter indicators suggest that media is increasingly inclined to take sides, and is increasingly viewing AI negatively.
An extensive survey conducted by CARMA (6,300 respondents in 19 large media markets across the world) showed, however, that the assumption that media both shapes and reflects public sentiment may have broken down to some extent when the subject is AI. Positive public sentiment toward AI outweighs negative views, and the mismatch between media messaging and public opinion is most obvious when comparing the “emotional signals” in reporting with specific feelings such as trust, excitement, fear and concern, anger or indifference expressed by survey respondents. According to the survey, people have somewhat more trust and excitement, and far less fear and concern and anger than what is portrayed in the media.
One important conclusion that CARMA drew from its study is that public interest has clearly shifted from interest in the technological novelty of AI to interest in its utility and practical value, which seems to be a consequence of personal experience with AI rather than influence from news headlines. Although the Philippines was not part of the survey, it would seem that would be reflected here as well; Filipinos are among the world’s most prodigious users of AI, according to recent studies, with more than 42 percent of internet users and up to 78 percent of workers reporting using AI in some fashion, and Metro Manila alone accounting for 2.5 percent of ChatGPT’s global traffic. The public sentiment, while overwhelmingly positive, is not uniform, however. People in industries such as manufacturing, health care, education and finance, where AI is perceived as labor-saving and augmenting human expertise, tend to have more positive views of AI, while sectors where AI is perceived as competing with people, such as the creative industries, media and some parts of the public sector tend to have more negative perceptions.
The main takeaway from all of this is that for policymakers and AI advocates, positive public sentiments and acceptance can be encouraged by transparency and focusing on the safe utility of AI. For those of us in the media, given that public trust in media has been shown to be declining generally, perhaps extra care should be taken when reporting on AI topics to ensure that the information we provide the public is indeed informative rather than argumentative.
AI potentially presents tremendous opportunities, but it is fraught with risks; that is the nature of transformative technology. Whether one’s role is to report on developments, form policy to take advantage of the opportunities and minimize the risks, or determine what part, if any, AI can play in one’s own business or daily life, knowledge about AI that is as accurate and complete as possible is critical.
