This morning, received an email from Academia.edu informing me that it had created an AI-generated podcast about an article that I authored a few years ago entitled, Representing Law Faculty Scholarly Impact: Strategies for Improving Citation Metrics Accuracy and Promoting Scholarly Visibility. If you’re not familiar with Academia.edu, it’s a commercial platform for sharing academic research.
I had the option of whether to include it on my Academia profile. I chose to add it since it represents an emerging trend in AI-generated content and scholarly visibility, two of my favorite subjects. This development reflects the growing use of AI tools to synthesize and present content, including scholarly work, in new formats, from podcast summaries to visual abstracts and interactive presentations.
Review of Academia.edu’s AI-Generated Podcast
The Academia.edu podcast did a poor job of synthesizing my article. It missed most of the major themes and instead picked up only a few scattered points to discuss.
The podcast also engaged in shameless self-promotion. At one point, it mentions that faculty can share their work on social media to increase visibility—something I did recommend in the article. Then it adds the example of “sharing drafts of papers on platforms like, oh, you know, Academia.edu—a little plug there,” which I did not say in my article.
Another aspect that troubled me was the lack of listener transparency about AI generation. The email I received as the author was upfront that the podcast was AI-generated. However, when you listen to the podcast itself, there’s no indication that it was generated by AI. Instead, it identifies the host as Richard Price, CEO of Academia.edu. I found this quite misleading.
Not all AI-Generated Podcasts are Created Equal: Enter Notebook LM
Before dismissing AI-generated podcasts of scholarship entirely, it’s worth noting that some tools handle this task much better. Google’s NotebookLM, which I recently blogged about as a document analysis tool, provides a striking contrast.
I asked NotebookLM to generate an audio overview of the same article and it did a very good job. It captured all the major themes and explored them in an engaging, coherent manner. The NotebookLM version demonstrated a much better understanding of the article’s structure and arguments, presenting them in a way that would be genuinely useful to listeners interested in the topic.
This difference highlights an important consideration for legal professionals and academics: as generative AI tools become more prevalent, we need to evaluate their quality and transparency carefully. GenAI shows promise for making scholarly work more accessible, but implementing it ethically and effectively matters.