A new international study challenges “one-size-fits-all” approach to AI adoption, revealing early users fall into four distinct archetypes with conflicting motivations and risk calculations 

As the global generative AI market is forecasted to surge to more than $440 billion by 2031, companies are rushing to adopt tools such as ChatGPT, Claude, and Gemini into their operations and offerings. Yet new research shows that a “one-size-fits-all” approach does not account for the various user groups of this technology. 

A study published in Electronic Markets – The International Journal on Networked Business by a team including researchers from a strategic partnership between the University of Oxford and the Berlin University Alliance reveals that the first wave of AI users is more complex than previously thought. The study, titled “Who uses general-purpose AI? A typology of ChatGPT early adopters,” identifies four distinct user archetypes, each driven by different motivations, needs, and priorities. 

Drawing on a two-stage survey of 344 early users conducted within the first four months after ChatGPT’s public release on 30 November 2022, Christoph Gerling of the Alexander von Humboldt Institute for Internet and Society (HIIG) in Berlin, Professor Timm Teubner of the Technical University Berlin and Dr Fabian Braesemann from the Oxford Internet Institute (OII) challenge the assumption that early adopters are monolithic group of tech-savvy enthusiasts – blinded by hype. 

The four faces of AI adoption 

By analysing participants’ responses related to behaviours, perceptions, and attitudes toward ChatGPT, the researchers discovered that adoption is driven by a complex mix of perceived utility, trust, privacy concerns and social presence (“how human it feels”). 

The study breaks down users into four specific archetypes: 

• AI Enthusiasts: The only truly tech-savvy group. They are highly engaged and seek both productivity and social benefits of modern-day AI. 

• Naïve Pragmatists: Utility-driven users who value convenience and results above all else. Yet they trust AI in ways that may expose them to risk. 

• Cautious Adopters: Curious and pragmatic, yet vigilant. These users constantly weigh the functional benefits against potential drawbacks. 

• Reserved Explorers: Sceptical users who are “dipping a toe in” but remain unsure of the technology’s benefits. 

Figure 1 – Aspects related to AI adoption. The figure illustrates that functionality alone is not enough to understand AI adoption patterns. Early adoption is driven by both functional and social-relational aspects. The extent to which AI feels “human-like”, its trustworthiness and privacy matter as much as productivity gains for many users. 

The privacy paradox 

One of the study’s findings is the prevalence of a “privacy paradox.” Three out of four identified user groups expressed significant privacy concerns, yet they continued to use the tools. This suggests users perform a fragile mental calculus, trading personal data for productivity – an issue made more pressing by the fact that 70% of non-users in 2025 cite data protection as a primary concern. 

Based on this finding, the researchers warn that anthropomorphising AI – making it seem more human – could backfire. Privacy-conscious users may begin to blame the AI itself for potential violations rather than the company behind it, eroding trust faster. 

Beyond traditional technology adoption theories 

The research demonstrates that established technology acceptance models are no longer sufficient to explain the adoption behaviour around generative AI. Unlike earlier chatbots or voice assistants, large language models enable open-ended, co-creative interactions that shift the user’s role from operator to curator. 

The study’s first author, Christoph Gerling (HIIG), explains: 

“Natural language interfaces have changed what ‘ease of use’ means. Using AI feels intuitive, but mastering it requires exploration, prompting skills and learning through experimentation. This makes the ‘task-technology fit’ more dependent on the individual than ever before.” 

In other words, how people use AI matters far more than whether they use it at all. 

Why one size fails to understand AI adoption 

As AI tools rapidly evolve from novelty to commodity, understanding user diversity is critical. Developers and businesses using a “one-size-fits-all” approach risk alienating major segments. The study argues that while analytical capabilities might sell a tool (or a service based on it) to an AI Enthusiast, trust-building and clear privacy safeguards are essential to win over the Cautious Adopter and Reserved Explorer. 

The authors conclude that understanding early adopters is crucial because their usage patterns reveal both the strengths and weaknesses of the new technology, offering a roadmap for wider adoption. Companies that recognise and address the nuanced needs of these diverse users are best positioned to capture disproportionate value in the Fourth Industrial Revolution. 

Download the paper, ‘Who uses general‑purpose AI? A typology of ChatGPT early adopters’.

Media spokespeople 

First author: Christoph Gerling, PhD Researcher, Alexander von Humboldt Institute for Internet and Society, Berlin 

Second author: Professor Timm Teubner, Technical University of Berlin 

Third author: Dr Fabian Braesemann, Departmental Research Lecturer, Oxford Internet Institute, University of Oxford 

Contact 

For more information or to arrange an interview, please contact:
Anthea Milnes, Head of Communications
Sara Spinks / Veena McCoole, Media and Communications Manager 

T: +44 (0)1865 287237

M: +44 (0)7551 345493 

E: press@oii.ox.ac.uk 

About the Oxford Internet Institute (OII)     

The Oxford Internet Institute (OII) has been at the forefront of exploring the human impact of emerging technologies for 25 years. As a multidisciplinary research and teaching department, we bring together scholars and students from diverse fields to examine the opportunities and challenges posed by transformative innovations such as artificial intelligence, large language models, machine learning, digital platforms, and autonomous agents. 

About the University of Oxford    

Oxford University was placed number one in the Times Higher Education World University Rankings for the tenth year running in 2025. At the heart of this success are the twin-pillars of our ground-breaking research and innovation and our distinctive educational offer. Oxford is world-famous for research and teaching excellence and home to some of the most talented people from across the globe. 

Funding information 

This research was partially funded by the Friedrich Naumann Foundation for Freedom with support from the German Federal Ministry of Research, Technology, and Space (BMFTR).