Dr Leah Henrickson writes for AI & Society.
As artificial intelligence continues to reshape how written content is produced, a new study examines how society perceives authorship in the era of AI-generated text—and whether those perceptions have shifted since before the rise of large language models (LLMs).
In Revisiting computer authorship: a longitudinal perspective, researchers Leah Henrickson and Leo Leppänen explore public attitudes toward authorship by comparing responses to an international survey conducted in 2024/2025 with an earlier version of the same survey from 2017/2018.
Key findings
- Authorship still feels human-centred: Despite dramatic advances in AI text generation, the study found only minor change in how people attribute authorship to computer-generated texts over the past decade. Many respondents continue to see authorship as tied to human identity and intentional creation, not machines.
- Ambivalence remains strong: Participants often expressed uncertainty about whether AI systems could be considered “authors” at all. This ambivalence reflects deeper cultural assumptions about creativity, responsibility, and human agency that have endured even as technologies evolve.
- Context matters: When survey participants were given more information about how a text was produced—such as whether a bot or a human created it—their likelihood of assigning human authorship changed. Yet, overall, clear consensus on AI authorship remains elusive.
Why this matters
As AI tools become more integrated into writing, publishing, and creative work, questions about *who gets credit—and who is responsible—*grow increasingly urgent. These findings suggest that public perceptions have not kept pace with technological capability, and that long-standing ideas about authorship still shape how we judge computer-generated content.
