I am a writer. Always have been. I actually have many fans of my writing. For sure I have critics. It comes with the territory. My english teachers/professors rotated between gifted and you'll never be a good writer. I am glad I ignored all of them.

This weekend (thanks to Howard Horn) I decided to reflect on my current usage of AI. In his article Howard advises that "reflection" is highly underrated and a crucial tool for us.. I agree.

So far I have two metaphors for AI in my writing life --

1. There is fast food, and then there is the sit down dinner  -- AI is fast food
2. Additional GPS measuring device I use on golf course -- confirmation confidence

When I golf -- before each shot I check multiple parameters

• the cart GPS (usually to back of green)
• my handheld GPS
• the embedded marker in the turf
• environmental factors (temperature, wind, elevation, etc)
• And finally -- I stand over the ball with the selected club and "eyeball" it and see what my gut tells me.

AI is much like Fast Food and very much like redundant GPS measurements. All theoretically good, but never the whole story..

I use AI quite often now but I use it much like my handheld GPS in golf but only as a secondary additional confirmation which can increase my certainty. And sometimes the cart GPS is marked wrong. What I see and feel is always topmost.

Recently, I experimented with two articles that Bassam wrote for me on service. Wonderful tone and ideal insight. I ran them both thru AI as "preprocessor" and compared the output.

The AI versions aim for brief bullet outlining but lose the nuance. Conditional outcomes are largely not presented in order to raise presumed clarity.

You can see for yourself in the two posts where I first use AI preprocessed and then the literal baseline article it was derived from.

• #2 -- SMB Selection framework for kiosks Part Two -- Whether you want a nice line item outline checklist, or you like the deeper examination, this article is for you. https://kioskindustry.org/part-two-smb-selection-service/
• #1 -- SMB Selection framework for kiosks Part One -- -- Whether you want a nice line item outline checklist, or you like the deeper examination, this article is for you. https://kioskindustry.org/part-one-smb-service-selection-framework-for-kiosks/

I am fond of saying there is no solution that fits all. Situations are different.

My audience is roughly 45/45/10

• 45% want the powerpoint bullet convention (probably run Windows too :-)
• 45% are interested in insight and obtaining their own basis for interpreting what they see
• 10% are the AI's trying to learn how to be smart and wondering how high I should rank. No different than Altavista and Yahoo for me back in late nineties when I thought being ranked on internet might be important. (I should relate funny stories about others keyword stuffing with mispelled phrases. e.,g. public computing turned into pubic computing.)

And I used to have to just debate people regarding content. Now I have to debate AIs.

Usual Caution (and this one is from AI)

There is growing evidence that heavy, passive use of AI tools can weaken everyday critical thinking, but deliberate, active use can also support and even improve it in some contexts.

What research is finding

Several recent studies and reviews report a negative correlation between frequent AI tool use and performance on critical-thinking measures, especially when people lean on AI for answers instead of reasoning themselves. This effect is strongest for younger, high-frequency users, while people with more education tend to better maintain critical-thinking skills even when using AI regularly.

Researchers describe this as “cognitive offloading”: shifting tasks like recall, analysis, or decision-making to technology instead of doing the mental work personally. When offloading is constant and unreflective, people show reduced reflective problem-solving, weaker information evaluation, and lower confidence in their own judgment.

Specific risks to critical thinking

Common risks that studies and educational analyses highlight include:

• Reduced practice: If AI routinely supplies explanations, arguments, or code, users get fewer reps at analyzing, synthesizing, and debugging their own thinking, so those skills atrophy over time.
• Automation bias: People tend to over-trust AI outputs, accepting them with little scrutiny and missing errors or biases, which undermines habits of questioning and verification.
• Cognitive overload and mental fatigue: Constant interaction with AI and large volumes of generated content can lead to information overload and diminished attention, which also hurts higher-order thinking.

Potential benefits when used well

On the other hand, AI can scaffold critical thinking when treated as a partner rather than an oracle. In learning and professional settings, decision-support and tutoring systems can surface alternatives, highlight trade-offs, or provide quick feedback, which can deepen understanding when users still must justify and choose among options.

Educational work suggests AI can enhance higher-order thinking if tasks require students to critique, compare, or improve AI’s output instead of merely submitting it. Some frameworks emphasize information literacy: people who are trained to check sources, question assumptions, and test AI-generated claims seem less vulnerable to erosion of critical-thinking skills.

Practical ways to protect your thinking

Experts generally recommend practices like:

• Use AI to generate options, then pause and decide independently, writing down personal reasoning before reading its suggestions.
• Routinely challenge AI answers: ask “why,” request alternative viewpoints, and verify important claims with primary or high-quality secondary sources.
• Reserve some tasks (mental math, planning, reading original material) for “no-AI” work to keep those cognitive muscles active.
• In teaching or self-study, structure prompts so AI is a sparring partner: have it critique your argument, then you critique its critique.

In short, AI use does appear capable of dulling natural critical thinking if it replaces rather than augments human reasoning, but thoughtful, constrained use with explicit habits of skepticism and independent analysis can mitigate that risk and sometimes even sharpen critical skills.