Media hype or scepticism shapes AI’s future as much as the technology itself

A research team at Corvinus examined how press narratives interact with public visions of AI’s future. The study’s lead author is Lilla Vicsek, Professor at Corvinus, and the article was published in Global Media and Communication in January this year. 

The researchers argue that the development of new technologies is often accompanied by periods of inflated expectations – so-called hype phases – followed by cycles of disappointment that also play out in media coverage. This recurring pattern is frequently described in the literature as alternating “summers” and “winters”. 

“Many people feel as if ChatGPT appeared overnight, but in reality it broke through at a moment when media attention towards AI was rising again,” Vicsek said. She added that generative AI was technologically mature well before 2022, yet the press paid relatively limited attention to it. 

More upbeat coverage when only industry voices are heard 

As part of the study, the researchers analysed AI-related articles from three leading US daily newspapers – The Wall Street Journal, The Washington Post and The New York Times – from the year before generative AI’s breakthrough. Across most of the coverage, the tone was restrained and often sceptical, emphasising AI’s limitations and unfulfilled promises. 

Topics such as mass job losses, “superintelligence”, or artificial general intelligence appeared only rarely in the articles analysed. When these concepts did surface, they were typically presented as unrealistic or exaggerated visions of the future rather than near-term possibilities. 

Among the newspapers studied, The New York Times showed the strongest critical approach, while the other outlets varied in how much they included independent or critical perspectives. The authors also highlight a clear pattern: when an article relied only on a company representative speaking about their own product – without other voices included – the overall tone tended to become noticeably more positive about AI. 

“It is important to recognise exaggerated technological promises in the media, because predictions about AI’s future are often made by people with a direct stake in the technology’s success. Balanced coverage requires journalism built on diverse sources — not only business and tech actors, but also civil society organisations, policymakers and critical researchers. The future of AI is not purely a technological matter; it is also shaped by social interpretations and media narratives,” Vicsek emphasised. 

Understanding these collective biases around technology matters because they can influence investment decisions, regulatory approaches and political priorities. In this process, the press plays a key role, especially high-prestige outlets that can shape how decision-makers think about what AI is, and what it might become.

A research team at Corvinus examined how press narratives interact with public visions of AI’s future. The study’s lead author is Lilla Vicsek, Professor at Corvinus, and the article was published in Global Media and Communication in January this year. 

The researchers argue that the development of new technologies is often accompanied by periods of inflated expectations – so-called hype phases – followed by cycles of disappointment that also play out in media coverage. This recurring pattern is frequently described in the literature as alternating “summers” and “winters”. 

“Many people feel as if ChatGPT appeared overnight, but in reality it broke through at a moment when media attention towards AI was rising again,” Vicsek said. She added that generative AI was technologically mature well before 2022, yet the press paid relatively limited attention to it. 

More upbeat coverage when only industry voices are heard 

As part of the study, the researchers analysed AI-related articles from three leading US daily newspapers – The Wall Street Journal, The Washington Post and The New York Times – from the year before generative AI’s breakthrough. Across most of the coverage, the tone was restrained and often sceptical, emphasising AI’s limitations and unfulfilled promises. 

Topics such as mass job losses, “superintelligence”, or artificial general intelligence appeared only rarely in the articles analysed. When these concepts did surface, they were typically presented as unrealistic or exaggerated visions of the future rather than near-term possibilities. 

Among the newspapers studied, The New York Times showed the strongest critical approach, while the other outlets varied in how much they included independent or critical perspectives. The authors also highlight a clear pattern: when an article relied only on a company representative speaking about their own product – without other voices included – the overall tone tended to become noticeably more positive about AI. 

“It is important to recognise exaggerated technological promises in the media, because predictions about AI’s future are often made by people with a direct stake in the technology’s success. Balanced coverage requires journalism built on diverse sources — not only business and tech actors, but also civil society organisations, policymakers and critical researchers. The future of AI is not purely a technological matter; it is also shaped by social interpretations and media narratives,” Vicsek emphasised. 

Understanding these collective biases around technology matters because they can influence investment decisions, regulatory approaches and political priorities. In this process, the press plays a key role, especially high-prestige outlets that can shape how decision-makers think about what AI is, and what it might become.