FAMILY9 - Confessions of a Blogger
I’ve been writing my blog now since July 2018, over seven years - a total of 159 articles. For the last two and a half months, eight articles, I’ve been using artificial intelligence (AI) in my research and to help write the articles. In this blog, I’d like to share with you my experience using AI.
First of all, I have long been
interested in the exciting, evolving AI technology. I even wrote a blog on the subject a couple
of years ago! So, when it became
available to help research and write articles, I was eager to try it out.
The bottom line is that AI was
useful beyond my wildest dreams. It
allowed me to expand my research efficiently, and actually helped me write the
articles!
But, I confess, that I am
suffering some unease at this new AI-assisted approach. With AI, all it takes is a quick prompt and
it can output a thousand words on any topic in seconds. It comes down to crossing the line between
being a writer and a copy-and-paste artist:
the difference between creativity and productivity. The AI product, while often impressive, tends
to sound like a Wikipedia article; it’s not me in the sense of my unique
writing voice that incorporates my personal experience, expertise, and
knowledge.
This article will cover a
snapshot of my blog writing experience without AI, then cover how I’ve used AI
lately, next summarize what I think are the benefits of AI in this context,
then my issues with using AI for writing, and close with a statement of my
blogger’s dilemma.
My Blog Creation Experience
without AI
Typically, once I choose a
subject, I Google that subject, and read a half dozen sources to get a feel for
the subject. Then I list the elements
that I want to cover in the article, prepare an outline, and look for key online
figures that I might use. Then I
consider the source material for each element on the outline and write a draft
for each outline element. This step
usually identifies questions I have, or concepts that I need to describe
better, which leads me to additional research to strengthen my draft. Having a reasonable first draft of the entire
article, I then scour the internet for images that support my work.
After a final edit, I ask Pat to
read and comment on the draft, after which I usually have more work to do
(especially explaining things better) before posting the article.
My Blog Creation Experience
with AI
After two years of preparation
(including providing AI overviews), Google rolled out “AI Mode” to everyone in
the U.S. on 20 May 2025. As Google in Ai
Mode says:
“AI Mode is an AI-powered search experience, designed to
handle complex and multi-layered questions. It leverages Google's AI
models, particularly Gemini, to
understand user intent, synthesize information from various web sources, and
provide comprehensive, AI-generated answers. It also allows for deeper
exploration through follow-up questions and links to relevant web pages.”
The figure below is a computer
screen shot from Google in Ai Mode, with a response to my query on what is “the
history of virtual reality,” the title and subject of my last blog.
Example computer screen shot for Google in AI Mode.
On the left of the screen is the
start of the Google AI Mode response which goes on for several pages. On the right of the screen shot is a list of
19 sources used for the response, with a link to each source.
I
will use “the history of virtual reality (VR)” as an example of my blog writing
experience with AI.
The AI response to
this initial query was substantial and enabled me quickly to get insight into
the key elements of the subject. I
explored the sources listed on the right of the above screen shot to get
additional information and to satisfy myself that the AI response was
accurate.
Using the initial AI response,
plus some of the sources, I made an outline of the article.
I then used the Google AI Mode to
explore the major elements on my outline.
In this case, I sought more information in each phase of the development
of virtual reality and for each major element of my outline, including current
applications of VR, challenges to its evolution, and the future VR. I also used AI mode to get a better
understanding of technical terms, add examples, and to expand on selected
items.
Then I put it all together,
sometimes using the AI-provide text, supplemented with additions/expansions of
my own creation.
And finally, of course, I had Pat
read the article and comment on it, which as usual necessitated some better
explanations of selected parts.
Benefits with AI
Google’s AI Mode biggest benefit
to me was in the research phase, providing an initial “kick start” in
identifying the key elements of the history of VR (some of which I hadn’t even
thought of), and a comprehensive list of sources for me to explore. Subsequent uses of AI Mode let me bore down
into specific aspects of VR. This saved me
a lot of time and effort.
I have not discovered any
errors in my AI Mode blog queries, but I have found errors in personal history
subject responses.
The AI Mode responses were for
the most part well organized and well written.
For broader queries like “AI development since the year 2000,” the
response contained logical/appropriate subheadings, making it easy to
follow. Often, the responses started
with a very well written overview and concluded with a summary.
I thought that AI Mode, when
asked, simplified complex subjects.
Overall, I found Google AI Mode
to be very efficient and useful, and it certainly saved me a lot of time
compared to researching/writing without AI.
Issues with AI
The biggest issue I found using
AI Mode, was that responses differed when the same query was asked multiple
times - not drastically, but certainly noticeably, including both the content
and organization of the response. This
bothered me, raising doubt about getting the “best,” most complete, and most
useful response. I found myself
comparing the different responses to the same question to satisfy myself that I
was getting everything I could.
Google AI Mode explains that
there is a degree of randomness in query answers based on probabilities of word
sequences when AI model settings allow for more variety, comparing different
responses to conversations with humans that aren’t always exactly alike. Also, AI models interpret context of the
query in slightly different ways with each repeated query, influenced by subtle
nuances in data processing.
AI Mode responses do not always
include supporting figures, and give you a rather sparse and incomplete group
of figures.
And as suggested earlier, the AI
Mode product, while often impressive, tends to sound like a Wikipedia article;
it’s not me in the sense of my unique writing voice that incorporates my
personal experience, expertise, and knowledge.
Blogger’s Dilemma
Google in AI Mode summarizes the blogger’s dilemma:
“Using AI in blogging presents a dilemma for many bloggers. AI tools offer advantages like faster content
creation, help with research and outlining …. However, they also come with
drawbacks, such as the potential for generic, inaccurate, or biased content, a
lack of the human touch that fosters engagement, and ethical concerns around
originality and plagiarism.”
And for the “Navigating the dilemma” query, AI
Mode suggests that: “The consensus among many marketers and content creators is
that a hybrid approach is the most effective way to leverage AI in
blogging. Instead of using AI to replace
human writers, it can be employed as a powerful tool to assist in various
stages of content creation,” including “brainstorming and outlining,” “research
and information gathering,” and “drafting and rewriting.” But, for
“editing and refinement: Human writers should meticulously edit AI-generated
content to ensure accuracy, eliminate bias, and infuse the human element that
connects with readers.”
And finally, AI Mode cautions that “Bloggers
should be transparent with their audience about the use of AI in content
creation, particularly when AI has significantly contributed to a blog post or
article.”
I agree with
this, and I
COULDN’T HAVE SAID IT BETTER MYSELF!
Comments
Post a Comment