Background:
In my T&E work domain, the adoption of AI tools has been encouraged to facilitate faster delivery of results.
Since the topics I write about on the MyGolfMk7.com blog overlap with my ‘real world‘ work, I started to see opportunities to apply AI tools to help create the website’s content.
Data:
Extracting data of interest from log files was my first use of AI tools to speed up data analysis tasks.
Then, leveraging the tools to analyze the data and generate charts showed good results.
Proof-reading:
Spell-checking tools are so ubiquitous that I doubt anyone gives a second thought to the help they provide in fixing incorrectly spelled words. Adding the Grammarly extension to my browser to help check grammar, spelling, punctuation, clarity, and tone was a big help in cleaning up my writing and ensuring complete sentences.
For a long time, I have sent selected draft posts to fellow enthusiasts or vendors for proofreading before publication.
This process of having the entire blog post reviewed by a second set of eyes was a good candidate for AI assistance. A person might not be available when I need help, and will certainly take longer to provide feedback. I also limited how often I asked for help, trying to be respectful and not ask too much of others.
Development:
The next evolutionary step was loading the AI tool with relevant references to help analyze advertising and flag misinformation.
The Federal Trade Commission’s Truth in Advertising and Substantiation Guidelines. Carnegie Mellon’s Argument Decomposition course. The Debunking Handbook (2020 edition). All of these are good references for evaluating advertising against.
I realized I now had access to a powerful weapon in the battle with Brandolini’s Law.
Brandolini’s Law:
Brandolini’s Law states that: “The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it.“
Also referred to as the Bullshit Asymmetry Principle, this means that if someone is spreading misinformation, refuting it will require at least 10x as much effort. The result is that debunking misinformation is an uphill battle.
The potential for more efficiently addressing misinformation with the help of AI tools is an exciting prospect.
Backlog:
The number of posts I’ve written down an idea of, and then not completed, is daunting. I have almost 200 draft posts that are waiting to be worked on. For some context on how much that is, this blog has been online for roughly 8 years (as of this posting), and I’ve published approximately 800 posts.
800 posts in 8 years works out to an average of 100 posts per year, and I’m almost 200 posts “behind“. In theory, I could publish posts for the next two years without needing to talk about anything I’m doing now. Nothing about the ongoing testing of the CC Design exhaust, the alternative turbine wheel for the IS48 turbocharger, the MGM7 intake, or fact-checking information from other sources.
I consider myself a good candidate to leverage AI support in more rapidly creating content for this blog.
It makes mistakes:
Although the AI tools are very helpful, they are still tools. If you’ve used one of the AI Tools, you might have noticed these cautions about relying on the tool output without checking for correctness.
I experienced this early on when I asked ChatGPT about a common topic that shows up on Mk7 discussion groups, recirculating Diverter Valves (DV) versus vent-to-atmosphere (VTA) blow-off valves (BOV).
I was just looking for a concise, factual summary of any performance differences, but what I got was the same “wrong answer” I see repeated in many discussion groups. What followed was a lengthy exchange between me and ChatGPT, which eventually led the AI tool to figure out why its original answer was wrong.
Lesson learned: take the developer’s note about it making mistakes seriously, and remain skeptical of AI-generated solutions.
With that being said, having extensive expertise with most of the topics this blog covers, and now being supported by an AI tool, opens up the possibility of covering much more ground and whittling away at some of the advantages that spreaders of misinformation have.
Next Steps:
I want to be upfront that I use AI tools to assist with various aspects of this blog’s content creation, as described above.
What AI doesn’t do is the core work this site is built on. All testing, measurements, analysis (with AI support), and conclusions come from my own hands‑on work. AI doesn’t replace my technical experience.
When I use AI for drafting or refinement, I provide explicit guidelines for how the tool should present information. Those guidelines emphasize objectivity, evidence‑based statements, and the minimization of emotional or personal language.
The goal is to keep the focus on data, scientific principles, and verifiable claims—not on personalities or assumptions about people.
These tools help me work more efficiently so I can spend more time running tests, validating claims, and sharing my findings with others.
Will I let AI take on more of the effort when fighting someone spreading misinformation or bullshit? Oh hell yes!
References:
- Federal Trade Commission – Advertising Guidelines for Businesses
- Carnegie Mellon University – Open Learning Initiative
- The Debunking Handbook (2020)