AI profiles you

In 2009 some people in academic management roles at my university underwent a 360 degree feedback process. I went through the process. A trained counsellor sought the views of a hand picked cohort of students, colleagues and other managers on my communication, leadership and other characteristics. She then reported back to me. As I recall I found it a very moving experience … as well as helpful.

I felt something similar in subjecting the contents of this entire blog to AI analysis. I started this weekly blog in 2010 and have since published 741 posts. I initially harvested material from previous publications, but later found that the blog posts could be assembled and repackaged to stimulate the production of a further series of academic books.

Today I saved the entire content of the blog as a text-only single PDF. The file comes to 30 Mb, about 787,000 words, i.e. 1,462 pages. I uploaded the file to Google’s NotebookLM platform (https://notebooklm.google.com/) that deploys ChatGPT-style large language modelling (LLM) tech to enable me, or anyone else, to interrogate the file.

The platform provided a helpful summary: “This document is a collection of biog posts by Richard Coyne, an architecture scholar. The posts explore the intersection of technology, culture, and design. Coyne uses his expertise in architecture, semiotics, and philosophy to analyse contemporary issues, from the impact of social media on human emotion to the ethics of digital money and the rise of the sharing economy. He examines how technology shapes our perceptions of reality, the role of language in shaping our understanding of the world, and the enduring relevance of classical philosophical concepts in the digital age.”

Personal information

As an initial test, I wanted to see what the model had to say about the author (me). It could not identify my gender from the blog content, or where I was located geographically. But when I asked about my age and career stage it answered: “While the sources don’t explicitly state the author’s age, there are clues that suggest they are likely to be middle-aged or older.” It cited examples from the apparent range of my knowledge, “understanding of complex philosophical and cultural theories, spanning from classical thinkers like Plato and Aristotle to contemporary figures like Jacques Derrida and Slavoj Žižek” which suggested “a substantial period of learning.”

It inferred from my mention of compiling book indexes that I had a significant period of academic experience, and my references to earlier technologies, such as typewriters suggested that I had been around when these were commonplace! It also identified specific blog posts where I seemed to offer advice to research students.

I asked if was able to identify my personality traits, specifically using the Myers Briggs personality model, which is normally assessed via a questionnaire about preferences. The AI provided a detailed assessment, along with the summary: “Overall, based on the sources, the author’s personality might align with an INTJ or INTP type. These types are known for their intellectual curiosity, analytical abilities, and enjoyment of exploring complex systems and ideas.” This outcome concurs with the results from various (MBTI) questionnaires I have completed in the past.

As for many of its responses, the platform provided the caveat: “it is important to emphasize that this is merely speculation based on limited information. Determining someone’s MBTI type accurately requires a comprehensive assessment, which is not possible based solely on a blog.”

My purchase list

I inquired about the content of the blog posts, which I will report on later, but I wanted to test how the AI was able to derive my potential purchase interests, i.e. how advertises might target me. After all, my blog is in the public domain, along with the blogs and social media posts of millions of other people — there to be mined and exploited. Here’s the full response from NotebookLM.

Products of Interest to the Author

Based on the sources and our conversation history, several product categories align with the author’s interests and potential needs:

Technology and Digital Media

• E-readers and tablets: The author frequently discusses e-books and digital reading experiences, referencing devices like the iPad and Kindle [1-3]. This suggests they are comfortable with digital reading formats and might be interested in newer models with advanced features or exploring different brands like Kobo or Onyx Boox.
• Electroencephalography (EEG) Devices: Source 9 mentions the Emotiv EPOC neuroheadset and its use in mapping emotional responses in the brain. The author’s interest in mood, cognition, and the relationship between technology and human experience could lead them to explore newer or more advanced EEG devices for personal use or research purposes.
• 3D Printers: The author discusses 3D printing and its implications for design, manufacturing, and consumer culture in source 30. They might be interested in inspecting or purchasing a 3D printer for personal projects, exploring the technology’s capabilities, or even researching its impact on design practices.
• Software for Writing and Research: The author’s prolific writing career suggests a need for robust and efficient tools for writing, research, and digital organization. Products of interest could include:
• Citation management software: Zotero, Mendeley, or EndNote for organizing and managing research references.
• Writing and editing software: Scrivener, Ulysses, or similar programs designed for long-form writing projects.
• Note-taking and mind-mapping software: Tools like Evernote, Notion, or Obsidian for capturing ideas, organizing notes, and outlining projects.

Books and Publications

The author’s extensive engagement with academic literature suggests an ongoing interest in acquiring new books and publications:

• Books on Philosophy, Semiotics, and Architecture: The sources reveal a deep interest in these fields, with frequent references to philosophers like Heidegger, Derrida, and Peirce. The author would likely be interested in new publications or classic works by these authors or exploring related areas like phenomenology, hermeneutics, and aesthetics.
• Books on Technology, Digital Culture, and the Internet: The author’s blog posts explore the intersection of technology, culture, and human experience. They might be interested in books examining the impact of digital media, social networks, artificial intelligence, or the philosophical implications of technological advancements.
• Subscriptions to Academic Journals: Access to relevant academic journals is crucial for staying up-to-date with research and scholarship. The author might be interested in subscriptions to journals focusing on architecture, design, philosophy, or technology studies.

Travel and Cultural Experiences

• Train Tickets and Travel Guides: The author’s description of a train journey from Sopron to the UK [1] suggests a preference for this mode of transport. They might be interested in exploring new train routes, particularly in Europe, and acquiring travel guides for destinations like Hungary or Austria.
• Tickets to Cultural Events: The author mentions attending the Edinburgh Fringe Festival [4], indicating an appreciation for live performances and cultural events. They might be interested in attending theatre productions, music concerts, or art exhibitions in their local area or during travels.

These product categories reflect the author’s intellectual pursuits, professional interests, and hobbies. Their specific purchasing decisions would ultimately depend on individual preferences, budget, and current needs.

So far, the results of my conversations with NotebookLM are impressive and on point. The platform deploys the Transformer model of language processing, and yet is able to process blocks of text in excess of the usual context window constraints.

Some may find the ability of the AI model to make such inferences disturbing. At the moment I think this technology is fascinating and requires further investigation.

Note

• Featured image is my ChatGPT: “Please generate an image to illustrate my blog post. An old style Freudian couch and chair ready to receive patient and analysis. Don’t show any people. Place a smoking pipe and ashtray within view. Show a reclining couch as found in a psychiatrist’s clinic. The analyst would sit in an upright chair behind the head of the patient who lies down.”

.