Jacek Manko - AI narratives - between superintelligence and power [Hype Studies Conference 2025 presentation]

Since its inception in the 1950s, artificial intelligence (AI) has been marked by ambiguity about its goals. Initially, AI aimed to computationally simulate intelligent behaviours, like playing logical games, Yet AI founders like John McCarthy and Marvin Minsky boldly envisioned "fully intelligent machines" or ones with human-level intelligence. Today, even though AI focuses on statistical data processing via machine learning, yet bold visions persist. OpenAI, with its 2022 ChatGPT release, revived narratives of AI "surpassing humans," referencing vague concepts like artificial general intelligence (AGI) and "superintelligence," which loosely suggest technology exceeding human intellect.

The study aims to examine contemporary forms of narratives regarding both the current state and further development of AI. Specifically, the study sought answers to the following questions:

• How do the largest companies in the AI industry define AI?
• How do they envision the further development of AI?
• Does the way AI is defined translate into the way these companies operate on a business level?

To this end, a qualitative content analysis method supported by virtual ethnography was applied. The analysis covered press releases, charters, or other public documents from major AI companies such as OpenAI, Anthropic or DeepMind.

Preliminary findings suggest that AI companies often invoke visions of "AI surpassing human intellect" yet fail to clarify what this means. This vagueness is intentional, as the focus is not on the technology itself, but primarily on power. Therefore, such arguments appear to be used instrumentally. That is, they serve as tools - initially for grabbing symbolic and imaginative power by shaping a specific vision of an AI-determined future. This, in turn, facilitates the acquisition of tangible power, such as securing investment funding and/or favorable political regulations.