Recently, I realized that Google Search has started cooperating with ChatGPT through an AI mode that delivers more organized and concise answers than the traditional “fragmented search” results. However, when I searched for my newly published paper online today, I was surprised — instead of retrieving the full article, the AI-generated summary presented the key points of my paper almost verbatim. This experience made me question the hidden flow of information behind these AI systems. It seems that beyond the surface-level promise of “smart answers” or “green business,” there might be deeper mechanisms at play regarding how content is accessed, processed, and reproduced.
The implications are concerning:
- Intellectual property may become effectively open-access before authors give explicit consent.
- Scientific knowledge is being distributed more rapidly — but possibly without proper attribution.
- Legal accountability becomes complicated, as AI-generated summaries often rephrase or slightly alter original texts to avoid straightforward plagiarism detection.
I’m still in shock, and honestly, more hesitant about sharing my creative or research ideas with ChatGPT or similar AI tools. It feels unsettling to think that these platforms not only charge users for access, but might also quietly benefit from our intellectual contributions — repackaging and reselling our ideas in the process. More discussion are needed.
I am thinking about writing a paper to reveal the phenomenon…
Maybe I can find some researchers to do this together…
Model: The Hidden Information Flow (red thread) in ChatGPT (Source: Authors’ own work)
RoofGazers 