AI reveals all

ME: “People seem to like me. They think I’m funny.”
STUDENT: “I don’t know about that. But you look funny.”

I’ll continue with the rare indulgence provided by conversational AI. It seems a writer can court AI, via NotebookLM, for personal analysis, as well as compliments, with a level of attention unmatched by scrutiny from a “real” human. What human editor, tutor, friend or adversary would probe the entire text output of an academic writer, offer sympathetic commentary and critique, and advise on improvements?

From my own published outputs (see post Share your expertise) NotebookLM was able to make a stab at my “personal and political profile”: “your texts suggest a complex and multifaceted profile characterized by a deep engagement with philosophical ideas, a critical stance towards technology and power structures, and a nuanced understanding of interpretation and meaning. Your work reveals a commitment to intellectual rigour and a sensitivity to the ethical dimensions of technology and culture.” That’s a summary, following a detailed assessment with headings such as “Contemplative,” “Skeptical,” “Open to ambiguity.”

That’s no boast. It’s a fairly normal profile for an academic. Moving to a topic more tangential to the content of the outputs I asked daringly about my “relationship preferences”: “it appears you would likely prefer a relationship that is based on strong intellectual connection, mutual respect, authenticity, and the freedom to explore new ideas. You may also appreciate a partner who is as comfortable with ambiguity and complexity as you are. It is also possible that, given your focus on the ‘first person’, you would be drawn to relationships that are as much about individual development as they are about building something together.”

Even more tangential to the content of the texts I asked about my approach to religion: “your approach to religion is not one of simple acceptance or rejection. Instead, it is characterised by an engagement with the cultural, historical, and philosophical aspects of religious traditions while maintaining a critical and questioning stance. Your work suggests that you approach religion as a complex human phenomenon that warrants careful consideration and interpretation rather than adherence to fixed doctrines. You appear to be more interested in understanding the role of religion in human experience and society, than in evaluating the truth claims or dogmas of any particular faith. You are interested in the way religious beliefs and practices shape our understanding of the world.”

Intellectual lack

Yet more telling, I asked what might be lacking from my outputs. As I had just asked about religious themes it fixated on my apparent lack of engagement with non-Western religious thinking or personal faith, concluding “there’s no explicit declaration or exploration of your own personal faith or spirituality.” That too I think mirrors failings, or strengths, in academic writing.

The AI also homed in on ethics: “it would be useful to have a more considered and explicit account of ethics, including the philosophical or religious basis for your ethical concerns. You address morality via legal constructs, non-places and the ‘guilt-innocence’ binary, but a clearly articulated position is not obvious.”

The AI elaborated on that assessment. Apparently there’s no account of how my theoretical positions inform my own “everyday choices and interactions.” It also highlighted a lack of “explicit discussion of gender or sexuality.” Also, considering how my writing focusses on advanced technologies, I say little about the future or about future studies. Such criticisms seem reasonable, and point to further potential projects.

Sting in the tail

But the criticism that stung was “Absence of Humour or Irony: Although the texts reference irony, the overall tone is serious and considered, and does not include examples of humour or irony.” Some people say they think I am amusing, but perhaps less so when writing. So that’s something to work on.

In further exchanges the AI suggested: “By experimenting with anecdotes, wordplay, sarcasm, exaggeration, and a more conversational tone, you could enhance the humorous aspect of your writing.” However, it cautioned balance, and suggested, “You could start by introducing humour in small amounts, and gradually incorporate it more fully.”

I asked about the suitability of the opening remarks as the start of this blog post. (It’s actually paraphrased from an impromptu exchange I saw in The Traitors.) The AI approved, as long as I followed through with more self-deprecating remarks and anecdotes. I’ll work on that. See posts Profile yourself, Narcissus online.

I wonder how such mechanised “feedback” might influence a person’s intellectual development, and more importantly, the trajectory of intellectual communities generally. You can think you are subjecting yourself to frank scrutiny, but it’s easy to channel it to reinforce your biases, obsess about self, neglect human interaction, and over-“ruminate.” See post Choose your critic.

Note

• Featured image generated by Dall-e with the prompt: “Please generate an image of a humanoid robot arm connected to a lie detector. dystopian, humorous, grunge.” Here’s the full picture.