An interesting conversation with ChatGPT on how it assists humans in earning intellectual ownership:
AI intelligence software are tools and cannot be authors because they cannot defend or revise work independently; they have no intent and no accountability. AI should not do all the thinking through low-level prompting.
In the discussion below too, AI assists in gathering requirements and examples; as the human author, I choose the lenses and tone, set the boundaries, and evaluate the points.
1. Authorship Requirements
Across journals, publishers, and ethics bodies (e.g., ICMJE — International Committee of Medical Journal Editors, COPE — Committee on Publication Ethics, APA — American Psychological Association), authorship requires:
- Substantial intellectual contribution
(conceptualization, argument framing, interpretation) - Critical decision-making
(what to include or exclude, tone, structure, emphasis) - Accountability
(ability to defend the work, responsibility for errors) - Original intent
(purpose, values, direction)
2. High-Level Prompting and Intellectual Ownership
High-level prompting with AI is required to earn intellectual ownership. This involves framing concepts with clear criteria, as well as critiquing and comparing argumentative points. The human defines the framework, chooses the lenses, sets boundaries, and evaluates outputs, while AI functions only as a cognitive assistant rather than a creator or author.
Examples:
- “Frame laziness as a moral versus systemic issue using biblical and psychological lenses.”
- “Critique this argument and rebuild it for a Christian Malaysian audience.”
- “Compare different religious interpretations and expose underlying assumptions.”
3. Clarification of ChatGPT’s Role
ChatGPT’s role includes language generation, pattern recognition, structural suggestions, and cross-domain synthesis. It qualifies as a tool similar to statistical software, grammar checkers, or reference managers. ChatGPT is trained on a mixture of licensed data, human-created text, and publicly available text.
It does not browse the internet unless explicitly enabled, and it does not recall sources unless they are cited by the user or accessed via tools. ChatGPT constructs language as output, but factual verification remains the responsibility of the human author.
Conclusion
AI functions much like a statistician: it recognizes patterns and constructs language, applies the rules of grammar, and synthesizes sources and requirements to help manage references. However, humans must still possess—or be willing to develop—the ability to verify data, understand statistical logic, evaluate language accuracy, and critically assess synthesized content. Intellectual ownership ultimately rests with the human, not the tool.
As a human author, I often use ChatGPT and Gemini to identify structural patterns in subjects that interest and logically sound to me, check my grammar, and provide references. Gemini also helps me generate images; its artistic skill is far greater than my own.