A new tool from Google allows for quick analysis of large amounts of text. Columnist Eric Thomas tried it out this week. (Eric Thomas illustration for Kansas Reflector)
This week, I met with a few dozen young journalists to talk about my work as a columnist for Kansas Reflector. During the presentation, I talked about my weekly writing process, my idea file and my writing quirks.
Each time that I present to an audience like this one, I get a bit reflective. What key issues in the state of Kansas am I ignoring in my writing? What political leanings are evident through these writings? How could I do better? Those questions send me paging through the 118 columns that I have written since 2021.
Even as the writer of those pieces, it’s difficult to find patterns. A typical column is 1,000 words, so all of those columns bound together are the rough equivalent of a 200-page book. This week’s reflective mood combined with such a large set of writing sent me in a new direction: to a new artificial intelligence research and writing tool.
Notebook LM was developed by Google Labs with the marketing line: “Go from information to insight in a snap.” In my week of trials with it, the marketing rings true. A few features are incredible and precocious, even in this moment when artificial intelligence is accelerating.
The goal of the app is to provide writers a way to create their own data sets to explore. You might provide the app with a dozen academic studies and ask it to summarize. You might upload all of your lecture and textbook notes from a chemistry course and ask for it to teach you a difficult concept in a new way. Or, you might upload a set of documents — for instance, 118 Kansas Reflector columns that you have written — and ask questions. So, that’s what I did. Upload and question.
What key issues in the state of Kansas am I ignoring? Notebook LM suggested that I focus more on health care and economics.
What political leanings are evident through these writings? “The writings lean politically left” because I concentrate on social justice and climate change as I cite sources such as the New York Times, the Atlantic and NPR.
I also asked about my writing style to test its analytical abilities. It scanned the dozens of columns and noticed my penchant for parentheses, colons and dashes — guilty as charged.
I am bringing news of this AI tool to you not necessarily because you are also a writer (although I hope you are, and if so, that you will start writing for Kansas Reflector). The significance of this app — and others like it — is that it uses artificial intelligence in more specific and practical ways than the large language models that dominate the news.
Artificial intelligence, in this way, is poised to emerge as a daily work companion rather than a parlor trick. Many of us have moved past the cutesy phase when we constantly asked Chat GPT if it loved us so that we could gauge its emotional guardrails. And sure, we can keep asking chat bots for anodyne advice, like recommending a travel itinerary for a road trip to the Smoky Mountains. The reality is that AI can be pretty bad at recreational trivia questions.
Notebook LM and other apps are different because of their powerful, niche uses in the workplace. Using this app has signaled to me the wide variety of narrowly tailored artificial intelligence products that we will likely use during our future work days.
Consider the variety of apps on your computer right now: one for video editing, one for writing email, one for spreadsheets and once for video conferencing. Many of those apps will continue to thread machine-learning into their existing functions.
It’s also likely that we will have a parade of new apps, ones that are almost exclusively oriented toward artificial intelligence. For me, that’s Notebook LM, an app that is there for an AI function, but that doesn’t replace my normal software for writing, design and teaching.
Besides giving me an effective critique of my columns, what else distinguishes Notebook LM?
Audio Overview: Notebook LM can produce a podcast-style summary of documents that you provide. I fed the app two studies about student attitudes toward free speech. After a few minutes of processing, two synthetic voices bantered and summarized for 12 minutes about what the related studies mean. For people like me, who gobble our news and knowledge with audio, the feature is equal parts magic trick and productivity boost. (It’s also an ego trip to hear voices, even if they aren’t “real,” discussing your work when you provide it.)
Citations: When you ask Notebook LM about the documents you have uploaded, it often answers the question with citations. It shows its work. If I want to know where the app found a specific statistic or evidence, I simply click on the footnote and it takes me to the original source material. With these citations, I worry much less about AI hallucinations.
Limited data set: Ask most artificial intelligence a question, and it might pull its answer from any of the billions of words that the large language model has ingested. Notebook LM, like many other apps, allows you to control and limit the data set that it is analyzing, so you aren’t basing business decisions on someone’s 2019 Reddit post. You are relying on the data that you trust.
Here’s one of the reasons why I found it so helpful to ask Notebook LM to review my work. Sometimes writing can feel a bit like tossing a message in a bottle into the sea. A handful of readers reach out every few columns, which is a blessing. However, as a person who thrives on feedback, I would love to know more.
This is my request to you, loyal reader who made it to the end of my column: How can my writing better serve you and our state? Please email info@kansasreflector.com and the editors will send your feedback along to me.
I ask, because that is a question that Notebook LM could not answer. And also because your answer, dear human, matters much more.
Eric Thomas teaches visual journalism and photojournalism at the William Allen White School of Journalism and Mass Communication at the University of Kansas in Lawrence. Through its opinion section, Kansas Reflector works to amplify the voices of people who are affected by public policies or excluded from public debate. Find information, including how to submit your own commentary, here .
Local News
